Harnessing the for Global Development

A CONTRIBUTION TO THE UN BROADBAND COMMISSION FOR SUSTAINABLE DEVELOPMENT

Table of Contents Forewords ...... 4 Mr . Houlin Zhao (ITU Secretary-General)...... 4 Mr . Chuck Robbins (CEO, Cisco Systems)...... 5 Executive Summary...... 6 Extending our Hyperconnected World through the IoT ...... 9 Defining the Internet of Things ...... 10 How the Internet of Things is Emerging...... 15 Functionalities of the Internet of Things...... 15 Functionality...... 15 Sensors...... 18 Connectivity...... 22 Applications across Different Sectors in Development ...... 24 The Internet of Things in a Developing Country Context ...... 25 Healthcare ...... 27 Water and Sanitation...... 30 Agriculture and Livelihoods...... 32 Resiliency, Climate Change and Pollution Mitigation...... 34 Natural Resource Management...... 36 Energy ...... 37 Other Sectors...... 38 Challenges to the Deployment, Impact and Scale of the IoT in Developing Countries ...... 41 Challenges of the Internet of Things...... 42 Technical Challenges ...... 43 Policy Considerations...... 43 Overlapping Issues...... 44 Recommendations – Supporting the IoT...... 46 Annexes...... 48 Annex 1: IoT Projects by Sector...... 49 Annex 2: The different characteristics of wireless IoT connectivity options...... 56 Annex 3: Sample sensor prices...... 57

1 Harnessing the Internet of Things for Global Development Figures

Figure 1: Miniaturizing & Multiplying – Getting Smaller & More Numerous (ITU, Mary Meeker, the Brookings Project)...... 12

Figure 2: The Intersection between IoT and (Cisco)...... 13

Figure 3: From Individuals to Society – Examples of data generated by the IoT (ITU)...... 14

Figure 4: Different IoT Applications with Different Characteristics (ITU)...... 16

Figure 5: Range of Common Sensors (Harbor Insights) ...... 19

Figure 6: Sensors by Retail Cost (Cisco)...... 21

Figure 7: Comparing IoT Connectivity Technologies (Cisco)...... 23

Figure 8: Access to Energy, Water & GSM Population Coverage in Sub-Saharan Africa (GSMA)...... 26

Figure 9: The Virtuous Circle of Development Impact (Cisco)...... 27

Figure 10: Areas of Highest Potential Impact across Different Sectors (ITU)...... 28

Figure 11: Monitoring the Movements of People during the Ebola Outbreak...... 30

Figure 12: Summary of Emerging Challenges in relation to the IoT (Cisco)...... 43

Tables

Table 1: Functionality within IoT Technologies (ITU)...... 17

Table 2: Sensor Types, Functionality, and Examples (Cisco)...... 20

Table 3: Examples of IoT interventions Mapped to the MDGs and SDGs (Cisco) ...... 39-40

Harnessing the Internet of Things for Global Development 2 Acknowledgements

This report has been written by Phillippa Biggs (ITU), John Garrity (Cisco), Connie LaSalle (Cisco) and Anna Polomska (ITU), under the supervision of Dr . Robert Pepper (Cisco) as a contribution to the ITU/ UNESCO Broadband Commission for Sustainable Development . Mitch Hulse and Mary Carol Madigan provided research support, with additional review provided by Dr . Ruba Borno (Cisco) .

The authors wish to thank the following people and organizations who generously contributed their time and insights through interviews and discussion to this report (listed alphabetically by organization/ project, and individual’s name):

●● American Red Cross (Abi Weaver)

●● DAI (Robert Ryan-Silva)

●● Echo Mobile (Zoe Cohen, Jeremy Gordon)

●● Integra (John Waugh, Eric White)

●● International Center for Theoretical Physics (Marco Zennaro)

●● Kopernik info. (Edwin Mulianto)

●● Network Startup Resource Center (Jonathan Brewer)

●● Nexleaf Analytics (Martin Lukac, Nithya Ramanathan, Shahrzad Yavari)

●● Oxford University - SmartPump Mobile Enabled Transmitter (Patrick Thomson)

●● Portland State University - SWEETSense, Inc . (Dexter Gauntlett)

●● Rainforest Connection (Dave Grenell)

●● United States Forest Service (Dr . Lindsey Rustad)

●● University of California Center for Effective Global Action (Temina Madon)

●● University of Washington - Smart Connect (Richard Anderson)

●● Wadi Drone project, NYU Abu Dhabi (Matt Karau, Martin Slosarik)

●● WellDone – MoMo (Austin McGee, Tim Burke)

●● World Agroforestry Center (Siddharth Mohan)

●● World Bank (Edward Anderson)

The views contained in this report may not necessarily reflect the views of the ITU or Cisco Systems .

*Front cover image: Achieving the Sustainable Development Goals (SDGs) with the Internet of Things

3 Harnessing the Internet of Things for Global Development Foreword by Mr . Houlin Zhao, ITU Secretary-General After more than a decade of discussion and anticipation, the Internet of Things is now firmly on its way . This is a major development, which promises to extend our online world in a myriad of ways . Connecting things, as well as people, offers prospects of new ways of monitoring situations, learning and responding in real-time .

In some sense, the “Internet of Things” is a misnomer . The Internet of Things is not a single, unified network of connected devices, but rather a set of different technologies which can be put to work in coordination together at the service and to the ultimate benefit of people in both developed and developing economies . This set of Internet of Things technologies is realizing a vision of a miniaturized, embedded, automated environment of devices communicating constantly and automatically . However, connecting up devices or robots (whether they are bridges, fridges or widgets) is only a means to an end — the really interesting part arises in terms of what can be done with the data obtained, and the learning outcomes for improving our future .

While researching this report, my staff learned it is not possible to focus solely on the technologies, at the risk of ignoring the human context in which these technologies must work . There are many difficult trade-offs involved— only some of which are technological (for example, the trade-off between robustness and reliability, and the sophisticated functionality of sensors on a water pump) . Other trade-offs enter into broader issues (for example, gaps between technical security and users’ perceptions of security and trust, or the detailed information yielded by geo-localization technologies) . Moreover, the purpose for which technology and applications are developed does not always end up as the sole — or even major — purpose for which they are actually used .

This report raises many important questions . Nevertheless, asking the right questions is also an important part of any learning process, and I welcome this report’s thoughtful exploration of the use and applications of some of these different technologies in the context of developing countries .

Houlin Zhao ITU Secretary-General

Harnessing the Internet of Things for Global Development 4 Foreword by Mr . Chuck Robbins, CEO, Cisco Systems At Cisco, we believe that technology has the capability to transform lives . Over the past few years, we have seen companies, cities and countries move to digitize all that they do, driving positive results in numerous ways . Yet across the world, many people still must confront a number of difficulties from climate change, endemic poverty, and environmental degradation, to the lack of access to quality of education, and communicable disease . While these big issues might seem insurmountable, we believe that technology can play a critical role addressing many of these challenges, while creating a host of opportunities .

Today, the Internet of Things is improving the day-to-day lives of citizens around the world . In cities from Barcelona to Songdo to Rio de Janeiro, Internet Protocol (IP)-connected sensors are monitoring traffic patterns, providing city managers with key data on how to improve operations and communicate transportation options . Similar information flows are improving hospitals and healthcare systems, education delivery, and basic government services such as safety, fire, and utilities . Sensors and actuators in manufacturing plants, mining operations, and oil fields are also helping to raise production, lower costs, and increase safety .

In both developed and developing countries, the Internet of Things is also helping monitor critical vaccines through the use of IP-connected thermometers . Moisture sensors in agricultural fields now alert farmers to the exact needs of food crops, and acoustic sensors in protected rainforests are helping to curb illegal logging .

For the IoT to have an even greater impact, there is still more we can do to improve the deployment of these technologies in developing countries . Network deployment, power requirements, reliability and durability are all uneven, at best, and policy considerations concerning access to data, legacy regulatory models, standards, interoperability, security, and privacy need to be addressed .

We are in the early stages of IoT adoption and are pleased to see the great impact that the IoT is already making for people, companies, industries and countries around the world . This report aims to serve as a guide for how the IoT can be fully utilized to improve the lives of people everywhere . The application of the IoT to solve many of the world’s challenges is limited only by our imagination .

Chuck Robbins CEO, Cisco Systems

5 Harnessing the Internet of Things for Global Development Executive Summary

Harnessing the Internet of Things for Global Development 6 This report explores the current use and potential of real-time and improved learning opportunities . IoT is Internet of Things (IoT) technologies in tackling global closely related to the concepts of Machine-to-Machine development challenges, highlighting a number of (M2M) communications and Wireless Sensor Networks specific instances where IoT interventions are helping (WSN) on the connectivity side, and to Big Data in to solve some of the world’s most pressing issues . It terms of the content outcomes produced . The IoT also presents summary conclusions on what is required for comprises the data produced and transmitted between the IoT to reach billions of people living in the developing machines (M2M), as well as between machines and world, and also to accelerate income growth and social people (M2P) . Key elements include machine-produced development as a result . data (e .g ., from sensors), and the communication of that data (via connectivity technologies) . Information and communication technologies (ICTs) such as mobile phones, Internet use and Big Data1 Many different stakeholders are involved in active IoT analytics are pervasively utilized in global development projects on the ground, including industry members, projects (in a field often known as ICT for Development, universities, NGOs, and tech start-ups, each contributing or ICT4D) to improve outcomes and deliver services . different strengths . The IoT is not just a story for Recently, this field has experienced strong growth . For industrialized economies or industrial applications, but example, in a recent presentation, Carolyn Woo, CEO is equally relevant for developing countries . The IoT of Catholic Relief Services (CRS), identified 157 new and connected sensors are driving improvements to development assistance projects, each started by CRS human wellbeing in healthcare, water, agriculture, natural between 2014 and 2015, that incorporate ICTs (primarily resource management, resiliency to climate change and mobile telephony2) . Similarly, at Johns Hopkins University energy (as reflected in the UN’s post-2015 sustainable alone, as of May 2015, there were over 140 mHealth development agenda) . The research for this report (mobile phone-enabled healthcare) projects across uncovered many interesting examples and applications the developing world .3 And an April 2015 review at the of the IoT in developed economies . However, these World Bank identified at least thirty-two projects that are beyond the scope of this report, which focuses on specifically incorporate Big Data analytics 4. As evidenced impactful applications of the IoT for developing countries . by these examples, mobiles are highly integrated in development projects already . Connected sensors and When determining which IoT application fits best M2M connectivity represent the next frontier in the for a particular context, there are many trade-offs ICT4D story . and compromises involved . Technical trade-offs include different characteristics among connectivity After having been coined as a term in 1999 by Kevin technologies, including, but not limited to: performance, Ashton,5 and after more than a decade of discussion and efficiency, reliability, robustness, flexibility, range, anticipation, the Internet of Things is finally emerging . power requirements, data throughput, cost (of sensors, This is a major development, which promises to change connectivity modules and service) and licensed our way of doing things through better information in versus unlicensed spectrum . For large-scale systems including hundreds of thousands of sensors, devices

1 “Big Data” refers broadly to a data set, comprised of structured or and/or readers, high reliability levels are likely to prove unstructured data, so large or complex that traditional data processing applications are not sufficient for analysis . important . Cultural context on the ground also matters,

2 Keynote presentation by Dr . Carolyn Woo of Catholic Relief Services http:// and it should be taken into account, along with technical schd ws/hosted_files/crsict4dconference2015a/84/2015%20ICT4D%20Confer. - ence%20Welcome%20Presentation%20Final .pdf considerations .

3 Keynote presentation by Dr . Alain Labrique of Johns Hopkins University at the Catholic Relief Services 2015 ICT4Development Conference, Chicago, May 27, Huge new opportunities are now opening up through 2015 . improved access to and use of Big Data techniques, 4 “Big Data for a More Resilient Future”, World Bank Group/ 2015 Spring Meet- ings, Big Data for Development, Summary of Projects . which offer learning opportunities to improve real-world 5 Kevin Ashton, “That ‘Internet of Things’ Thing, in the real world things matter processes and enhance decision-making over the more than ideas,” RFID Journal, June 22, 1999 .

7 Harnessing the Internet of Things for Global Development short-, medium- and long-term in healthcare, education, Maximizing the benefits of the IoT is likely to require more emergency services and disaster response, among a coordinated regulation across all sectors, with telecom/ variety of other application areas . ICT regulators working closely with their counterparts in data protection and competition, but also with Impactful IoT interventions in development can improve emergency services, health and highway authorities 6. efficiency (achieving similar levels of impact with fewer resources) and/or enhance effectiveness (increasing Laws and regulations on data will need to be impact with similar levels of existing resources) . In reconsidered carefully in view of the IoT — in terms advancing global development, IoT interventions are of how data are obtained and can be used, how long helping to improve research, public policy, basic service data can be kept, limits on access by third (fourth or delivery and the monitoring and evaluation of fifth… nth) parties (the term ‘third party’ may prove programmes across a range of different sectors . This woefully inadequate in some cases) . The information report discusses examples of use cases of the IoT in collected from sensor systems may or may not be freely healthcare, water, agriculture, natural resource accessible on the Internet (Open Data), and the data management, resiliency to climate change and energy . transmitted may or may not cross the public Internet .

The IoT has regulatory implications across the areas of Given the high pervasiveness of the IoT’s impact, it is vital licensing, spectrum management, standards, that as more countries introduce policy frameworks, they competition, security and privacy – only some of which take into account the various factors and implications of are the familiar territory of telecom regulators, compared the IoT across different sectors . When all stakeholders to other domains where non-telecom regulators may are included in active dialogue, the IoT represents a typically take the lead . promising opportunity for more coherent policy-making and implementation .

Photo credit: Marco Zennaro . GSM connectivity node in Nairobi .

6 “Regulation and the Internet of Things”, GSR-2015 Discussion Paper, available at: www .itu .int/en/ITU-D/Conferences/GSR/Documents/GSR2015/Discussion_ papers_and_Presentations/GSR_DiscussionPaper_IoT .pdf

Harnessing the Internet of Things for Global Development 8 Extending our Hyperconnected World through the IoT and Big Data

Photo credit: Marco Zennaro . Solar-powered, Wi-Fi enabled soil moisture node at the Asian Institute of Technology in Bangkok .

9 Harnessing the Internet of Things for Global Development Defining the Internet of Things For example, ABI Research separates the “digital-first” domain (e .g . PCs and mobile devices, designed mainly After more than a decade of debate, discussion and as digital devices) from the “physical-first” domain (e .g . anticipation, the ‘Internet of Things’ (IoT) is finally humans and other, hitherto unconnected “things”) . emerging . As early as 2005, the ITU noted that the ABI Research also notes that despite the ongoing development of the Internet of Things as a function convergence of these two distinct domains, there is still of our hyperconnected world encompassed a set of only one overarching Internet through which the domains technological advances from different fields — specifically, connect 5. A 2012 Machina-GSMA study distinguishes wireless and mobile connectivity, nanotechnology, “mobile-connected devices,” which connect directly to a radio-frequency identification (RFID) and smart sensor mobile network (usually via a SIM card) from the broader technologies 1. Advances in these technologies, when “connected devices” market, which includes “short- combined, could help realize a miniaturized, embedded, range devices” (e g. ., devices using Wi-Fi, Bluetooth and/ automated Internet of connected devices communicating or other connection technology) .6 regularly and relatively effortlessly .2 The UK regulator OFCOM differentiates between the two Today, governments, businesses, and consumers are related terms: using the IoT and Big Data to introduce new business models, to improve the delivery of services, to increase ●● M2M: describing the interconnection (usually via efficiency in production, and to enhance wellbeing wireless technologies) of devices that would not and human welfare . As with many other technologies, previously have had the ability to communicate (e .g ,. vendors, implementers, operators, policy-makers and connected cars) . regulators aim to maximize the benefits of deployment ●● The IoT: a broader term describing the interconnection while minimizing potential risks to security and privacy . of multiple M2M applications, often enabling the exchange of data across multiple industry sectors Widely disparate definitions of the IoT exist . The ITU (e g. ., the ability to manage traffic flow, reduce pollution has defined the IoT as “a global infrastructure for the and improve health by combining data from a range of information society, enabling advanced services by transport, healthcare and environmental sensors) . interconnecting (physical and virtual) things based on existing and evolving interoperable information and Deloitte sees M2M as a category that has become communication technologies” (Recommendation ITU-T broader over time, encompassing all types of telematics Y .2060) 3. The IoT clearly includes M2M (referring over cellular networks on trucks, smart utility meters, specifically to communication directly between eReaders, tablets and PC modems, but excluding 7 devices, used in a vast array of applications and for smartphones . They suggest that it is inappropriate for a variety of purposes4), but broader definitions of IoT M2M to include eReaders, tablets and PC modems technologies also include and smart since, “although there is the occasional automatic environments . data upload or download, most of the traffic via these devices is human-initiated and human-observed .” On this basis, Deloitte excludes smart TVs, games consoles 1 ITU (2005) . “ITU Internet Report 2005: The Internet of Things .” Available at: www .itu .int/osg/spu/publications/internetofthings/ and www .itu .int/osg/spu/publi- and set-top boxes from its predictions . The research cations/internetofthings/InternetofThings_summary .pdf . consultancy GP Bullhound agrees that it is the automatic 2 Opinion is divided on how many of these connected devices really need to interconnect and communicate with each other at any point of time . Some commentators suggest only 2% of devices need to communicate (for a specific purpose) . The ‘smart environment’ vision foresees nearly all devices within an 5 https://www .abiresearchcom/market-research/product/1017637-inter . - ‘environment’ communicating . net-of-everything-market-tracker/

3 ITU-T Recommendation Y .2060, note, s .8 .4 . 6 http://www .gsmacom/newsroom/press-release/gsma-announces-the-busi . - ness-impact-of-connected-devices-could-be-worth-us4-5-trillion-in-2020/ 4 For example, Vodafone has defined M2M as “remote wireless data inter- change between two or more devices or a central station that allows remote 7 Deloitte (2015) . “Technology, Media and Telecommunications Predictions monitoring and control of devices and processes .” See Vodafone press release 2015,” page 6 . Available at: http://www2 deloitte. com/content/dam/Deloitte/. on 29 April 2010 . “M2M : A new way of working .” Available at: http://enterprise . global/Documents/Technology-Media-Telecommunications/gx-tmt-pred15-full- vodafone com/insight_news/2010-04-29_a_new_way_of_working. .jsp report pdf.

Harnessing the Internet of Things for Global Development 10 initiation of traffic and data exchange that is the Comparable estimates include: distinguishing feature of M2M 8. 11 ●● 25 billion ‘networked devices’ by 2020 (ITU); The IoT is perhaps best understood as a set of related ●● 24 billion ‘connected devices’ by 2020 (Machina technologies that can be used together to achieve Research in conjunction with GSMA);12 and exciting ends, and it can be defined in terms of its contributing technologies, including the use of sensors, ●● 26 billion deployed IoT devices by 2020, a thirty-fold 13 RFID chips, nanotechnologies and identification systems increase from 2009 (Gartner) . (chips, cards, SIMs), among others . Overall, IoT and These broad estimates are considerably higher than various related technical developments (including Analysys Mason’s 2011 prediction that there will be convergence, cloud services, data analytics and the 2 .1 billion M2M device connections by 202014, which is proliferation of sensors) are resulting in: based on a narrower definition of M2M .

1. greater monitoring and measurement of humans, However, other sources place this figure considerably machines and things; as well as higher still, including the well-known Ericsson estimate 2. a shift from human-to-human communications of 50 billion connected devices by 2020, a figure with to M2M, something-to-everything, and everything- which the Hamilton Project (Figure 1, page 12) agrees, to-everything communications . after factoring in non-IP connected devices (such as RFID tags) . ABI Research estimates that the installed 3. Greater and more rapid awareness of and base of ‘active wireless connected devices’ exceeded 16 information about status, function, and environment . billion in 2014, up 20% from 2013, and that it will more than double to 40 .9 billion by 2020 15. These different definitions of M2M, IoT, and what constitutes a “network-enabled device” result in Based on their definition, Deloitte estimates that IoT widely varying estimates of the number of connected hardware could be worth around US$10 billion and IoT devices — depending on whether mobile, tablets, PCs, services worth around US$70 billion in 2015 alone 16. and wearables are included in definitions of ‘connected Anecdotal reports from various firms corroborate the devices’ . It is unclear whether many of these estimates market size and opportunity 17. ABI Research suggests include wearables . that IoT hardware and connectivity revenues are growing Gartner predicts that there will be 6 .4 billion connected at 10-20% annually, while apps, analytics and services things in use worldwide in 2016 9. However, in mid-2015, are growing at 40-50% annually . Based on rapid Cisco estimated that there were already 15 .7 billion growth, Gartner estimates that IoT product and service “devices connected to the Internet” – including mobile suppliers could generate incremental revenue in excess phones, parking meters, thermostats, cardiac monitors, 11 “ITU Predicts 25 Billion Networked Devices by 2020”, 28 September 2012, tires, roads, cars, supermarket shelves and many other http://www v3. co. .uk/v3-uk/news/2207590/itu-predicts-25-billion-networked- devices-by-2020, via ACM TechNews . types of objects .10 12 GSMA, quoted at: http://www .gsma com/newsroom/press-release/gsma-. announces-the-business-impact-of-connected-devices-could-be-worth-us4- 5-trillion-in-2020/

13 Gartner, Forecast: The Internet of Things, Worldwide, 2013, at https://www . gartner com/doc/2625419/forecast-internet-things-worldwide-.

14 S . Hilton, Imagine an M2M world with 2 .1 billion connected things… Analysys Mason, 27 January 2011 .

15 https://www .abiresearchcom/press/the-internet-of-things-will-drive-wire . - less-connect/ 8 GP Bullhound Technology Predictions 2015 . Available at: http://www .gpbull- hound com/wp-content/uploads/2015/01/GP-Bullhound-Technology-Predic. - 16 Page 6, Deloitte (2015), « Technology, Media and Telecommunications tions-2015 .pdf Predictions 2015”: http://www2 deloitte. com/content/dam/Deloitte/global/Docu. - ments/Technology-Media-Telecommunications/gx-tmt-pred15-full-report .pdf 9 http://www .gartner com/newsroom/id/3165317. 17 www .forbes com/sites/aarontilley/2015/05/15/qualcomm-the-internet-of-. 10 Cisco, Visual Networking Index 2015 . Available at: http://www cisco. com/c/. things-is-a-billion-dollar-business/; http://bits .blogs .nytimes com/2014/10/09/. en/us/solutions/service-provider/visual-networking-index-vni/index .html ge-opens-its-big-data-platform/?_r=0

11 Harnessing the Internet of Things for Global Development Figure 1: Miniaturizing & Multiplying – Getting Smaller & More Numerous

Each new computing cycle typically generates around 10x installed base of the previous cycle

Devices or users in millions; logarithmic scale

1,000,000 Mobile Internet 100,000 Desktop 10,000 Internet

1,000 PC 108+ Units

100 18+Units/Users Minicomputer 100MM+ 10 Units Mainframe 10MM+Units Devices / Users (MM in log scale) Devices 1 1MM+Units

1960 1970 1980 1990 2000 2010

The cost of computing power is falling rapidly

50 $10,100,000,000

Altair 8800 Number of Connected Devices (Billions)

$100,000,000 40 IBM PC

$1,000,000 30 $10,000 Gateway G6-200

20 2013 iPad2 >15 Billion (Log Scale - 2010 Dollars) (Log

$100 10 2010 2000 12 .5 Billion 1969 1984 200 Million 1,000 4 an iPad2 to Computing Power Cost of Equivalent 1960 1970 1990 2000 2010 2020

Sources: Mary Meeker’s Internet Trends Report 2014 (top), available at: http://qz com/214307/mary-meeker-2014-internet-. trends-report-all-the-slides/; The Hamilton Project, Brookings Institute, Ericsson (bottom) .

Harnessing the Internet of Things for Global Development 12 of US$300 billion by 2020, resulting in US$1 .9 trillion in information communicated to individuals (such as global economic added value 18. IDC forecasts that the automatic text messages from one’s bank regarding worldwide market for IoT solutions will grow from US$1 .9 account updates) . trillion in 2013 to US$7 .1 trillion in 202019 . M2M communications include machine-generated data The Internet of Things (comprising connected devices transmitted to other machines that process the data and connected environments, such as M2M and M2P) and determine which further actions may need to be and Big Data are separate, but related, concepts . There taken . For example, onboard vehicle sensors transmit is signifi cant overlap between these two concepts, data on engine, transmission and wheel performance to especially where data are collected and transmitted via a central processing capacity, which then determines connected devices and IP (Figure 2 below) . Person-to- at what point the vehicle may need to be serviced person (P2P) communications are direct interactions to prevent a breakdown . Simply put, the IoT involves between individuals, facilitated by IP, such as emails, machine-generated or machine-processed data that are video conferencing, SMS and phone calls . Machine-to- communicated over IP . person (M2P) transactions include machine-generated

Figure 2: The Intersection between IoT, M2M, Big Data and examples in international development

The Internet of Everything and Big Data Everyone

e .g .: Mobile money P2P e .g .: Tracking mobile signals transactions for population migration after epidemic outbreaks (e .g . e .g .: Hand water Cholera/Haiti; Ebola/Liberia) pumps equipped to send text message The Internet e .g .: Paper census records reports of faulty M2P digitized and then analyzed pumps of Things

e .g .: Networked M2M e .g .: Aggregate water pump smoke and fi re fl ow and use data analyzed sensors to transmit to inform new pump location warnings between decisions homes in densely populated informal e .g .: Aggregate data from fi res in settlements settlements is analyzed to inform urban planning decisions

Note: P2P: Person-to-Person; M2P: Machine-to-Person; M2M: Machine-to-Machine . Source: Cisco Systems .

18 www .gartner .com/newsroom/id/2684616

19 www .idc .com/getdoc .jsp?containerId=prUS24903114

13 Harnessing the Internet of Things for Global Development These sub-categories in Figure 2 correspond roughly and beyond can be improved in the developing world with ABI Research’s categorization of the “Internet through the implementation of IP-enabled sensors and of Everything”20 as a technology consisting of three actuators . Governments, development organizations, subsystems: businesses and citizens in the emerging world are already incorporating the IoT and Big Data analysis to ●● the Internet of Digital; help alleviate some of the developing world’s most

●● the Internet of Things (broadly corresponding with pressing problems . M2M in Figure 2); and M2M is often used for telemetry, robotics, status tracking, ●● the Internet of Humans (broadly corresponding with road traffic control, logistic services and telemedicine . Person-to-Person or P2P in Figure 2 above) . Vodafone notes that “M2M technology can be used for Another case in point is the purpose of IoT applications . […] a huge range of applications including data collection, These technologies benefit developing countries, in remote control, offsite diagnostics and maintenance, addition to their developed counterparts, in driving remote monitoring, status tracking, fleet management, social and economic advancement . Services ranging traffic control and security systems” 21. Figure 3 illustrates from water and sanitation to healthcare, agriculture

21 Vodafone press release, “Delivering M2M communications for the global 20 ABI Research definition available at: https://www .abiresearch com/mar. - business”, 30 April 2009, available from Vodafone press office at: http://enter- ket-research/product/1017637-internet-of-everything-market-tracker/ prise vodafone. com/insight_news/deliver_m2m. .jsp .

Figure 3: From Individuals to Society – Examples of data generated by the IoT

Individual Community Society

Level

Connected Cars Smart phones Smart Cities Health devices Wearables Smart Grids IoT Smart homes

Intelligent Transport Systems GPS, Fitbits Event Data Recorders (EDRs) Smart metering; Visa PayWave Blood pressure monitors; Smart water meters Examples Mastercard Paypass remote burglar/heating Traffic monitoring Employee passes systems

Speed, distance, airbag, Mobile money Electricity/water crash locations/alerts; Fitness data, GPS consumption & billing; Data Heart rate, blood pressure, location-based data Traffic flow data Diet, remote heating data

Intended Individual person GP, health authorities; Authorities/regulators Immediate friends/ family; health & car insurance; Utility companies; Audience banks; employers police, social networks Other citizens

Source: Adapted from the ITU Draft GSR Discussion Paper 2015, “Regulation and the Internet of Things”, Professor Ian Brown, Oxford Internet Institute, University of Oxford, UK . Note: Individual smartphones can be used to connect cars or control smart homes through tethering, so there is some crossover between individual and car/home connectivity .

Harnessing the Internet of Things for Global Development 14 how IoT applications at the individual level (including According to the ITU GSR 2015 Discussion Paper on GPS location, payment systems and wearables) can be regulation and the Internet of Things,25 one possible aggregated up into applications for smart homes and explanation for why the IoT is advancing rapidly now connected cars . At the macro level, three of the areas of is that it is moving from a position where it delivers greatest IoT development and investment are smart cities incremental efficiency improvements to existing business (where infrastructure and building systems could improve models to one where it positively impacts new business the efficiency and sustainability of a range of urban models and processes as well 26. activities), smart power, and water grids . In terms of existing processes, the IoT can improve Wearables are included in the definition of the IoT for the and enable a broad range of applications — from more purposes of this report, because most wearables either efficient manufacturing, logistics, counterfeit detection, communicate directly with another device (e .g . heart rate monitoring of people, stock, vehicles, equipment monitors that connect via Bluetooth to a smartphone) and infrastructure, to improved healthcare, traffic or they connect directly and upload data to the Internet . management, product development and hydrocarbon This may sound like a large number of applications, but exploration . according to the GSMA, pure M2M connectivity and In addition, the IoT is now also enabling the exploration sensor networks may represent only a small subset of of new business models such as car and truck rental some of the broader use cases for 5G connectivity 22. clubs, whose members can book and use vehicles How the Internet of Things is Emerging parked around their neighborhood almost on-demand; or “pay-as-you-drive” insurance based on driving patterns, The strong growth currently observed in IoT applications is behavior, and risk . For marketers, the IoT enables brands attributable to several major underlying trends that are just to gather more information about their customers, and now coming to fruition: create “truly compelling, magical experiences .”27

●● The reduction in the cost of computing (including Functionalities of the Internet of Things sensors) and the growth of Wi-Fi are enabling factors Any IoT system includes a set of functions and driving growth in IoT applications;23 parameters (Table 1, page 17), many of which are ●● Growth in mobile and the deployment of data-friendly common to other types of connectivity as well . IoT 3G networks from 2001 onwards, as well as the systems can be defined according to a set of criteria expansion of network connectivity across the world, and in terms of their communications technologies, range, from urban to rural settings (including Wi-Fi, but also necessary bandwidth, number of nodes, robustness, macro cell connectivity); reliability and/or power requirements (more often referred

●● The rise of software development, partly attributable to to simply as ‘battery life’) . economies of scale; and Functionality

●● The emergence of standardized low-power wireless For any IoT scheme, a number of trade-offs are involved technologies (as suggested by ABI Research24) . in performance, efficiency, reliability, robustness, flexibility, power supply, scalability, interoperability, ease

25 ITU Draft GSR Discussion Paper 2015, “Regulation and the Internet of Things”, Professor Ian Brown, Oxford Internet Institute, University of Oxford, UK, 22 GSMA Intelligence (December 2014), “Understanding 5G: Perspectives on at: www .itu .int/en/ITU-D/Conferences/GSR/Documents/GSR2015/Discussion_ future technological advancements in mobile . papers_and_Presentations/GSR_DiscussionPaper_IoT .pdf

23 “Internet of Things By The Numbers: Market Estimates And Forecasts”, 26 ITU Draft GSR Discussion Paper 2015, “Regulation and the Internet of Forbes, 22 August 2014, available at: http://www .forbes com/sites/gil. - Things”, Professor Ian Brown, Oxford Internet Institute, University of Oxford, UK, press/2014/08/22/internet-of-things-by-the-numbers-market-estimates-and- at: www .itu .int/en/ITU-D/Conferences/GSR/Documents/GSR2015/Discussion_ forecasts/ papers_and_Presentations/GSR_DiscussionPaper_IoT .pdf

24 https://www .abiresearchcom/market-research/product/1017637-inter . - 27 Andy Gilder . Beyond wearables: brands could capitalise on the Internet of net-of-everything-market-tracker/ Things . Memeburn, 6 June 2013 .

15 Harnessing the Internet of Things for Global Development of authentication, preservation of privacy, extensibility, sensors, devices and/or readers (e .g . for seismic mobility support, and modularity . For example, trade-offs monitoring networks around nuclear plants, but not may arise within the following: necessarily for RFID tags on livestock) .

●● Low latency (the time required for round-trip data ●● Complexity versus robustness transmission) becomes all the more vital for advanced (e .g . of sensors and connectivity modules); cloud computing applications (such as high-definition ●● Range, data throughput, and cost video conferencing and industrial collaboration), where (e .g . Wi-Fi vs LTE vs LoRa vs ethernet); any interruption or delay in data transmission can have 29 ●● Battery life versus data throughput . According to Cisco, major consequences . new models of electrical generation and transmission The choice of final specifications among the different (including batteries, simple chemical reactions, energy options available depends on the aims of the deployment, harvesting devices, etc .) will be needed to power the and the challenges to be solved . For example, Figure 4 28 multitude of new devices that will emerge . defines some of the technical characteristics of several See Annex 2 for discussion on different battery life different IoT applications . characteristics of different connectivity modules .

●● High reliability levels may prove very important in large-scale systems using hundreds of thousands of

28 Pepper, R . & Garrity, J . (2014) The Internet of Everything: How the Network 29 Pepper, R . & Garrity, J . (2014) The Internet of Everything: How the Network Unleashes the Benefits of Big Data . Global IT Report 2014 . WEF . http://www3 . Unleashes the Benefits of Big Data . Global IT Report 2014 . WEF . http://www3 . weforum org/docs/GITR/2014/GITR_Chapter1. .2_2014 .pdf weforum org/docs/GITR/2014/GITR_Chapter1. .2_2014 .pdf

Figure 4: Different IoT applications with Different Characteristics

Data Quality of Amount of Time Server initiated Packet Example Applications Mobility volume Service signaling sensitivity Communication switched only

Smart energy meters no yes yes Red charging yes no yes eCall yes no no Remote maintenance no yes yes Fleet management yes yes no Photo frames no yes yes Assets tracking yes yes no Mobile payments yes no yes Media synchronisation yes yes yes Surveillance cameras no yes yes Health monitoring yes yes yes

very low low intermediate high very high

Source: Handbook: Impacts of M2M Communications & Non-M2M Mobile Data Applications on Mobile Networks, page 50 . ITU (Geneva, 2012) . Available at: www .itu .int/md/T09-SG11-120611-TD-GEN-0844/en .

Harnessing the Internet of Things for Global Development 16 Table 1: Functionality within IoT Technologies

Functionality Examples

Fixed or mobile location (e .g . GPS, GPRS) or for monitoring the status of the engine in a 1. Location and mobility mobile car or on a flying drone .

2. Identification and addressing For example, IP addresses, IMEI, chips, smart cards, SIMs .

This refers to the degree of dispersion of nodes within an environment . There may be many close together (a dense environment) or few far apart . In a very densely connected 3. Topology, architecture and environment, an Internet of disparate connected devices may become a ubiquitous, fully node density and dispersion connected environment (e .g . a connected hospital, smart cities) . Topology takes into account the architecture and how nodes are constructed for reporting purposes (e .g . zone architecture or parent-child tree architecture with hierarchies) .

Detection/reporting of status . Many different types of sensors exist (e .g . thermal, moisture, optical, radiation, acoustic, kinetic, acoustic, chemical, position, instrumentation, motion, 4. Purpose level, GPS coordinates) for monitoring a range of variables associated with tracking position (e .g . weather, climate, water, light, salinity, soils, vegetation, etc .) - see Figure 5 overleaf .

Centralized or distributed communications via central or local servers, central analysis and processing . In many developing countries with available mobile networks, communication modules rely on wireless connectivity (GPRS, 3G, WiFi, WiMAX, etc .) . Data communications can be further broken down into other variables: 5. Efficient transmission/ ●● Range; exchange of information and ●● Routing: multi-hop or dynamic flooding; performance ●● Data volume and bandwidth needs – high or low; ●● One-way or two-way transmission capability; ●● The data may or may not be transmitted via the public Internet; ●● Latency and QoS of transmission .

Independently operated or back-up power system . The battery life may have an important 6. Power system trade-off with the weight and/or complexity of the sensor . Some devices will need (aka battery life) constant, wired power .

Some use cases for M2M require the connected device in the field to lie dormant for 7. Dormancy versus regularity periods of time, or to use ‘lean,’ regular, or ‘bursty’ signaling (to note, continuous transmis- of communication sion has zero dormancy) .

Commonly found in systems which alert to natural disaster . This may be binary (working/ 8. Alert/warning system fault; heart beat versus no heart beat) or set for the detection of motion (seismometer) or threshold (e .g . detection of pollution) .

Automation/independent function – whether human initiation, monitoring or intervention 9. Automation is required .

10. Frequency of visit/ For maintenance and/or obsolescence and replacement . replacement

Source: ITU .

17 Harnessing the Internet of Things for Global Development Sensors simultaneously applied to thousands of products on a manufacturing floor . Sensors can be broadly deployed to Much of the functionality of the IoT and the data overcome a host of challenges but, in some cases, they transmitted in the M2M and M2P notions of the IoT are may need to be highly customized . It is this customization determined by the nature of sensor measurement . Other that enables sensors to provide real added value for IoT machines associated with the IoT include actuators, and ICT4D initiatives . devices that can be directed to perform a physical activity such as opening an irrigation dam or closing a Appropriate ICT sensor technologies that address livestock fence . However, in the context of development, challenges surrounding primary healthcare, agriculture, the majority of current applications utilize connected aquaculture, water treatment, and air quality can be sensors . deployed in an efficient and cost-effective manner . Figure 5 (page 19) illustrates some commonly found As with the decline in the cost of computing power, sensors . the costs per unit of sensors have dropped steadily over time . Today, sensors are found in many everyday devices, and some of the latest smartphones come with at least ten embedded sensors, for example: a microphone to capture sounds, camera(s) to capture images (front and back), a fingerprint sensor, global positioning system (GPS), accelerometer, gyroscope, thermometer, pedometer, heart rate monitor, light sensor, touch screen, and barometer (not to mention the various connectivity technologies such as Wi-Fi, Bluetooth, GSM/ CDMA, LTE and NFC) 30.

Crop and livestock management serve as primary areas where connected sensors can provide data and facilitate ease of use for farmers and ranchers dealing with unpredictable climates . Sensor measurements also aid in the detection and prevention of major natural disasters . With pressure, temperature, and weather pattern sensor technology, appropriate emergency preparedness strategies can be used to brief, alert, educate and, if need be, evacuate populations in high-risk areas . Sensors equipped with accelerometer or seismographic capability can warn populations in high-density urban areas of impending earthquakes, tsunamis or typhoons .

Sensors are one of the primary modes of realizing the full potential of value added to companies, communities, and individuals that employ them for IoT purposes . In general, sensors host a heterogeneous school of functions . They can detect everything from changes in temperature and humidity to the amount of force and pressure being Photo credit: Marco Zennaro . Solar powered air quality sensor

30 http://www .phonearenacom/news/Did-you-know-how-many-different- . in Cotonou, Benin, sending data via SMS . kinds-of-sensors-go-inside-a-smartphone_id57885; http://web .stanford edu/. class/cs75n/Sensors .pdf

Harnessing the Internet of Things for Global Development 18 Figure 5: Range of common sensors

Position/ Presence/ Proximity Machine Vision/Optical Ambient Light Motion/Velocity/Displacement

Acceleration/Tilt Temperature

Electric/Magnetic Humidity/Moisture

Leaks/Levels Acoustic/Sound/Vibration

Force/Load/Torque Chemical/Gas Strain/Pressure Flow

Source: Harbor Insights .

Figure 5 highlights the range of common sensors and motor vehicles) to aid in anti-logging and available on the commercial market, while Table 2 anti-poaching efforts . (page 20) segments sensors by example and use case: ●● Chemical and gas sensors are used to measure oxygen, nitrates, pH, salinity, black carbon and carbon ●● Position, presence, proximity sensors such as the dioxide . Additional types of sensors can measure leaks Global Positioning System (GPS) and transponders can and levels of fluids and gases, or levels of electricity provide information about location . and magnetism which can provide information about ●● Motion, velocity, displacement include kinetic instrumentation such as battery life . sensors that can measure movement, e .g . for wind ●● Flow sensors are used in water pipes, pumps, bodies and water direction, speed, period, precipitation, of water, and air . barometric pressure and motion detection, as well

weight measurement (e .g . for animals or loads) . ●● Force/load/torque/strain/pressure sensors can be Accelerometers measure movement and gyroscopes used for weight measurements as well as in monitoring measure tilt, and digital vernier calipers can measure infrastructure (such as traffic flow over bridges) . the size of cracks . ●● Machine vision and optical ambient light sensors

●● Temperature sensor measurements are for ambient can measure different elements of visual bandwidth air or substances, such as water or soil . Similarly, (including color, intensity and/or clarity) . humidity/ moisture can measure the presence of air, ●● Radiation sensors can measure radiation levels and be water and other substances, including toxic gases . used remotely to measure sites for nuclear spills 31.

●● Acoustic, sound and vibration sensors can be used to identify wildlife (for research or monitoring) as well man-made noise (including for gunshots, chainsaws 31 Interview with John Waugh, April 13, 2015 .

19 Harnessing the Internet of Things for Global Development Table 2: Sensor Types, Functionality and Examples

Measurement Functionality Sensor Examples Use Cases

Land management; natural Detect and respond to RFID, linear position sensors, resource/wildlife Proximity/Position angular or linear position GPS position sensors, management; illegal activity of device location finding tracking

Ultrasonic proximity, optical Detect movement outside reflective sensors, Passive Emergency preparedness; Motion/Velocity/ of component within sensor infrared (PIR), inductive land management; illegal Displacement range proximity, accelerometers, activity tracking gyroscopes

Thermometers, resistant Water access; water Detect amount of heat in temperature detectors, treatment; agriculture; emer- Weather/Temperature different mediums and thermocouples, infrared gency preparedness; land metrics thermometers management

Detect decibel level sound or Seismography, firearm Emergency preparedness; Acoustic/Sound/Vibration seismic disturbances sensors, commercial security illegal activity tracking

Pressure monitors, capacitive Detect force(s) being exerted Natural resource Flex/Force/Pressure/Load transducers, piezoresistive against device management sensors, strain gauges

DC/AC electrical current Detect chemical, gas, or Agriculture; natural resource sensors, voltage transducers, Chemical/Gas/Electric electrical changes in management; health; smart home sensors, humidi- composition of substance water treatment ty monitoring

Detect color and light shifts Real-time temperature Light/Imaging/Machine Vision Health through digital signaling monitoring (infrared)

Source: “Roadmap for the Emerging Internet of Things”, Carre & Strauss; and Cisco authors’ adaptation .

Prototyping feasible IoT solutions is aided by the ubiquity and enhances innovation in this space . Low-cost sensor of sensor devices in production and the low cost of technology must be used in tandem with microprocessor entry for users . A complete state-of-the-art IoT sensor and circuit devices . suite can be deployed for approximately US$250 . In fact, the overall cost of sensor installment, deployment, At a minimum, a single sensor is usually deployed with a and support has decreased by a hundred times over single-board micro-controller computer device, such as 33 the past decade 32. This low cost of entry makes sensor a or board . technology highly accessible to aspiring IoT developers,

32 https://gigaom com/2015/01/25/declining-sen. - 33 http://sweden .nlembassy org/binaries/content/assets/postenweb/z/zweden/. sor-costs-open-up-new-consumer-applications/ netherlands-embassy-in-stockholm/iot_roadmap_final_draft_0309145 .pdf

Harnessing the Internet of Things for Global Development 20 These micro-controller devices serve as essential IoT Deploying sensors to meet ICT4D solutions can be purpose-built platforms for the end-user and often cost-effective . Realistically, any value gained from require the end-user to be familiar with software utilizing IoT sensor technology could be realized within engineering principles to achieve effective functionality . four to six months post-deployment . For example, a temperature, wind, and precipitation sensor suite in a Not all ICT stakeholders need access to full-fledged rural agricultural community could collect and analyze sensor suites to meet their solutions . Solution- weather data over a three-month timeframe that could appropriate sensors can be sourced and purchased for be used to better prepare the community for upcoming US$50-150 depending on the necessary functionality weather patterns and result in better crop and land needed to meet various ICT4D challenges, as noted in management moving forward . It is important to note that Figure 6 . the real value from using IoT-enabled sensors derives

Figure 6: Sensors by retail cost

Functionality Sensor Type

Highest Cost ●● Long-term install/deployment ●● Chemical/Gas

●● Industrial scale deployment ●● Electrical/Capacitive $150-$1000+ ●● Extreme accuracy/precision ●● Pressure/Load/Weight ●● Typically large enterprises ●● Proximity/Position

●● Ease of solution interoperability

●● Residential/commercial ●● Water Treatment/Flow

●● Advanced development kits ●● Weather/Temperature

●● Consumer-based support ●● Motion/Velocity $50-$150 ●● Cloud partnership capability ●● Acoustic/Sound/Vibration ●● Fast deployment ●● Light/Imaging

●● Medium infrastructure required ●● Proximity/Position

●● Low-Medium accuracy/Precision ●● Flex/Force/Strain

●● Single function ●● Water Treatment/Flow

●● DIY/Prototyping often needed ●● Weather/Temperature

●● Limited without other hardware ●● Motion/Velocity $0 - $50 ●● Requires basic equipment ●● Acoustic/Sound/Vibration ●● Geared towards amateurs ●● Light/Imaging

●● Singular functionality

●● No infrastructure required Lowest Cost

Source: Authors; see annex 3 for further details .

21 Harnessing the Internet of Things for Global Development from the ability of experts, analysts, and communities to Each technology has distinct characteristics, including use their collected data eff ectively and effi ciently to drive the range of their signal, the extent of their data current ICT-centered projects and to predict and plan throughput (or bandwidth), and the power needs of the for future challenges that could be addressed with IoT communications device (or battery life), among other technologies . attributes . Annex 2 discusses these in detail34 .

Connectivity ABI Research notes that technologies such as The connectivity requirements of diff erent types of IoT Bluetooth and ZigBee are helping drive node/sensor networks vary widely, depending on their purpose and implementations, while Wi-Fi or cellular technologies are 35 resource constraints . A range of diff erent wireless and providing the backbone for data transfer to the cloud . wireline technologies can be used to provide full IoT In terms of individual technologies: connectivity (Figure 7) . IoT devices communicate using ● Bluetooth is a high-speed, short-range and high data a range of diff erent communication protocols, which may rate but low battery life wireless technology that can include: short-range radio protocols (such as ZigBee, replace wired devices . Bluetooth and Wi-Fi); mobile networks; or longer-range radio protocols (such as LoRa) . These technologies can ● Radio protocols such as Ultra-Narrow Band can be segmented based on wireless versus wireline, and the provide longer-range coverage (useful for wireless technologies can be grouped by personal area applications such as video monitoring) . network (WPAN), wireless local area network (WLAN) or wide area network (WWAN) technologies .

34 See also http://www .eejournal .com/archives/articles/20150907-lpwa/

35 https://www .abiresearch .com/market-research/product/1017637-inter- net-of-everything-market-tracker/

Figure 7: Comparing IoT Connectivity Technologies

Wireless- Wireline Personal Area Wide Area Networks Local Area Networks (WLAN) Networks (WPAN) (WWAN)

ANT+, Bluetooth, 4 .0 LE LoRa, Weightless, Dash 7 Copper/DSL RFID, NFC Wi-Fi WiMax, 2G, 3G Coaxial 802 .11 .4, ZigBee 4G/LTE, Satellite Fiber

Range short to long

Bandwidth narrow to broad

Battery Life short to long

Note: Non-exhaustive Source: Cisco Systems . See also Annex 1 .

Harnessing the Internet of Things for Global Development 22 ●● Low-rate wireless personal area network (LR-WPAN) The FCC’s expert IoT Working Group has predicted that is a low-cost, low-power and low data rate wireless IoT will add significant load to existing Wi-Fi and 4G technology used for inexpensive mobile devices . mobile networks 39. Regulators need to give continuing attention to the availability of spectrum for short-range ●● Devices communicating over kilometres may access the IoT communications, and the capacity of backhaul 300 MHz to 3GHz spectrum range; networks, as well as encouraging the roll-out of small-cell ●● Centimetre or millimetre contactless transactions may technology and 4G . Assuming these conditions are met, use Near Field Communications (NFC) in 13 MHz or the Working Group does not expect that new spectrum Extremely High Frequency (EHF) bands . will need to be explicitly allocated to IoT communications .40

●● Some IoT applications may also make use of AM/FM Cisco observes that spectrum requirements include: bands in the Very High Frequency (VHF) range . narrowband and broadband frequencies; short-haul and long-haul spectrum; continuous data transmission and Until at least 2010, GSM remained the most widely used short bursts of data transmission; and licensed spectrum technology for M2M,36 especially in areas with coverage in addition to license-exempt spectrum . where advanced data transmission had been needed . However, given that many modern multi-sensor networks Ensuring device connectivity and sufficient bandwidth only require occasional connectivity with minimal for wireless sensors requires careful planning 41. It is the throughput and signaling load, GSM connectivity may not combination of these different qualities within Big Data be the most appropriate technology for certain projects . capacity that may offer the most exciting opportunities . For example, location information may be combined with When sensor networks send SMS at regular intervals, status information to provide real-time information on costs can accumulate very quickly . Depending on an evolving situation . There are huge new opportunities the nature of the sensor network and the terms and opening up by improved access to and use of Big Data conditions of the mobile contract, widespread use of techniques, which offer learning opportunities to improve SMS may require negotiation of a separate contract for real-world processes and enhance decision-making IoT traffic . For example, in one project in Indonesia, the over the short-to long-term in healthcare, education, mobile operator determined that SMS was being used for emergency services and disaster response . “business” rather than for “personal” use and changed the fee structure accordingly . There is a risk of bill shock and/ When integrated into an early warning system, real-time or lock-in with a particular provider or providers . data can be used to forecast potential outbreaks of violence or natural disaster . Text analysis of social media Wi-Fi is a short range wireless technology often used data has the potential to reflect growing tensions in a in mobile devices (e .g ., PDAs, laptops, tablets, etc .) . region given high unemployment or political frustrations 42. However, Wi-Fi is playing an increasingly important role in Light emission data collected via satellite-produced the IoT, with Wi-Fi chips embedded in portable computers remote sensing images can also be analyzed to estimate a and smartphones able to operate on a license-exempt country’s GDP in real-time 43. (unlicensed) basis37 and with the majority of the upgrade costs falling on consumers rather than operators . By 2011, one in ten people were using Wi-Fi, while today, Wi-Fi is 39 US FCC Technological Advisory Council IoT Working Group, Spectrum: Initial Findings, FCC TAC meeting update, 10 June 2014 . Available at: http://transition . in 25% of homes around the world, with two billion Wi-Fi fcc gov/bureaus/oet/tac/tacdocs/meeting61014/TACmeetingslides6-10-14. .pdf 38 devices sold in 2013 . 40 US FCC Technological Advisory Council IoT Working Group, Spectrum: Initial Findings, FCC TAC meeting update, 10 June 2014 . Available at: http://transition . fcc gov/bureaus/oet/tac/tacdocs/meeting61014/TACmeetingslides6-10-14. .pdf 36 World Bank Broadband Strategies Handbook (2011) . 41 Pepper, R . & Garrity, J . (2014) The Internet of Everything: How the Network 37 The ITU has designated the 2450 MHz and 5800 MHz bands for industrial, Unleashes the Benefits of Big Data . Global IT Report 2014 . WEF . http://www3 . scientific, and medical (ISM) applications that “must accept harmful weforum org/docs/GITR/2014/GITR_Chapter1. .2_2014 .pdf interferences .” This is often interpreted to mean that they are considered unregulated . See Frequently asked questions on the ITU-R website . 42 http://www .unglobalpulseorg/projects/can-social-media-min . - Available at: www .itu .int/ITU-R/terrestrial/faq/index .html#g013 . ing-add-depth-unemployment-statistics

38 Wi-Fi Alliance . Available at: www wi-fi. org/organization. .php 43 ITU (2014) . “Measuring the Information Society .”

23 Harnessing the Internet of Things for Global Development Applications across Different Sectors in Development

Photo credit: Juozas Cernius/American Red Cross . The photograph shows the installation of a smart home sensor network for fire detection in informal settlements in Nairobi, Kenya .

Harnessing the Internet of Things for Global Development 24 The Internet of Things in a Developing by using white space spectrum rather than other Country Context networks with high-speed capabilities, according to the GSMA,4 in areas where there is low interference from This section examines deployments of IoT technologies analogue broadcasting channels . in developing economies, covering what has worked and formulating summary conclusions on key issues to 4. Connectivity may begin with essential applications consider when extending the IoT to the billions of people only, which could be introduced initially on a small living in the developing world . The following specific scale and might not always become fully integrated considerations may apply: (e g. . as suggested by ‘greenshoots’ isolated pilots in eHealth and mHealth) . 1. Increasingly observed in developing countries, more of the population has access to basic 5. More constrained resources and fragile environments telecommunication network coverage than has may make populations in developing countries access to fundamental services such as electricity, inherently more vulnerable to natural disasters . running water and basic sewage facilities . The ITU For example, the Red Cross believes “emerging estimates that, in 2015, over 95% of the world’s technologies will play a particularly important role population resides within the coverage area of in amplifying efforts to facilitate community-level a 2G mobile-cellular network (and 69% under a knowledge and health, connection, organization, 3G network) 1. Figure 8 (page 26) compares and economic opportunities, access to infrastructure and contrasts urban and rural areas in sub-Saharan services, and management of natural resources”, Africa by their access to electricity, water and mobile and has convened a Global Dialogue on Emerging coverage . Technology for Emerging Needs of general relevance 5. 2. Economic sectors and processes in developing countries are more labor-intensive and may lack This does not mean that the developing world lacks supporting processes (e .g . agricultural systems may the prospects or potential to develop great applications not use technology-driven crop management, pest/ itself — just that deployments may differ from developed disease control or quality management systems) .2 country contexts in their purpose, resilience or supporting The Macrothink Institute notes that the information infrastructures . In developing countries, for example, requirements of productive systems in developing deployments may be more likely to be made in isolation countries are likely to be very different from those and independent of supporting infrastucture . The much- in developed countries, meaning that monitoring cited application of M-Pesa introduced by Safaricom systems for developing countries are likely to need in Kenya began essentially for unbanked consumers, different design requirements and compared with the development of wireless payment technological frameworks 3. systems in industrialized countries, where mobile money transfers may accompany or complement the formal 3. Lack of resources means that simpler, more banking system . cost-effective solutions may prove more effective in a developing country context . For example, Fixed-line infrastructure is also often more prevalent in using a wireless wide area network – WWAN – for developed country contexts, where IoT applications can communication could lower the cost of M2M modules run off high-capacity fixed line infrastructure, compared to developing countries, where IoT or sensor networks

1 “ITU Facts and Figures – The World in 2015 .” Available at: www .itu .int/en/ ITU-D/Statistics/Pages/facts/default .aspx 4 http://www .gsmacom/mobilefordevelopment/wp-content/uploads/2013/01/ . 2 Karim, Anpalagan, Nasser & Almhana (2013) . “Sensor-based M2M Sustainable-Energy-and-Water-Access-through-M2M-Connectivity .pdf Agriculture Monitoring Systems for Developing Countries: State and Challenges .” Macrothink Institute . 5 “A Vision for the Humanitarian Use of Emerging Technology for Emerging Needs: Strengthening Urban Resilience .” Available at: https://drive .google com/. 3 See Footnote 2 . file/d/0B1vf6TLGIC0yZk9UU2t2UGFmNlE/view?pli=1

25 Harnessing the Internet of Things for Global Development Figure 8: Access to Energy, Water & GSM Population Coverage in Sub-Saharan Africa

Rural Urban

60%

49% GSM Population Coverage

12% Electrification Rate User Improved Water Access

58%

83% Total Population 88%

Source: GSMA, “Sustainable Energy & Water Access through M2M Connectivity . http://www .gsma com/mobilefordevelopment/. wp-content/uploads/2013/01/Sustainable-Energy-and-Water-Access-through-M2M-Connectivity .pdf .

are most likely to be used in conjunction with mobile In developing country contexts, a range of uses infrastructure . and applications of IoT technologies first started to appear a decade ago . Impactful IoT interventions in IoT interventions are increasingly common in advanced development either improve efficiency (achieving similar economies . Wearable sensors in watches, bracelets and levels of impact with fewer resources) or effectiveness even clothes can help users monitor their vital signs and (increasing impact with similar levels of resources) 6. improve their health and wellbeing . In homes, offices IoT applications could help promote monitoring and and factories, sensors can detect when rooms are in and evaluation, and achievement of nearly all the existing out of use, enabling more efficient heating and lighting Millennium Development Goals (MDGs) and post-2015 and helping improve working conditions . In many more Sustainable Development Goals (SDGs) (Table 3, pages environments, smart meters can coordinate the energy 39-40) . For example, sensors in agricultural fields are consumption of appliances to smooth out variations in monitoring soil conditions and moisture levels . RFID overall energy consumption and achieve more effective tags are helping farmers provide more personalized use of variable renewable energy sources .

6 James Bon Tempo . Available at: http://linearityofexpectation .blogspot . com/2015/02/the-purpose-of-ict4d-in-one-diagram .html

Harnessing the Internet of Things for Global Development 26 Figure 9: The Virtuous Circle of Development Impact

Research

Monitoring Policy & Evaluation Formulation

Service Delivery

Source: Cisco Systems . care for their livestock . Connected thermometers are Healthcare monitoring vaccine delivery and storage in real-time . The IoT has the potential to improve health and wellbeing Cameras and sensors in smartphones and tablets are through greater efficiency and improved care in existing allowing healthcare workers to provide remote diagnosis healthcare settings, by enabling greater use of remote of disease . And off-grid solar systems, monitored via telehealth provision, and enabling individuals to monitor SMS, are bringing affordable electricity to lower income their own health day-to-day, improve wellbeing and families . better manage conditions (such as stress, encouraging In tackling global development challenges, IoT exercise and healthy eating), diagnose medical interventions are being utilized across the full spectrum conditions more quickly and promote treatment regimes . of development activities (Figure 9) . Figure 10 In terms of preventative care, IoT fitness devices such as (page 28) highlights how Big Data capacities enable both Fitbits, and movement sensors now built into many new real-time monitoring and response, and also predictive smartphones, enable many individuals to monitor and pattern data analysis, facilitating a shift in strategy track themselves, which generally promotes healthier from reactive to proactive . Academic researchers are lifestyles . For example, Apple and Google have added deploying sensors to improve research on increasing features to their latest smartphone operating systems to agricultural yields, for example . Public policy is being integrate health sensor devices and promote users to informed by data collection on community water usage . monitor their own health data using non-specialist health tracking apps .

27 Harnessing the Internet of Things for Global Development Figure 10: Areas of Highest Potential Impact across Different Sectors

Ex-post Current Future

Evaluation Measurement and Prediction and Assessment Real-time Feedback and Planning

Mobile Social Algorithmic Algorithmic Agent Enhanced Financial money network liquidity fraud detec- network credit agent analysis needs Services tion monitoring scoring placement marketing prediction

GDP Text analysis Text analysis Economic Income Mapping estimates Migration economic commodity and poverty social Develop- through monitoring downturn fluctuation assessment divides ment mobile data prediction prediction

Assessment Disease Migratory Predicting Health of mobility containment population outbreak restrictions targeting tracking spread

Geo-tar- Mobile data geted links Pests, bad Agricultural to track food between harvest yield/shock Agriculture assistance suppliers/ alerts predictions delivery purchasers

Social Predictive Social Campaign network algorithms network effective- delineated to anticipate Commercial targeted ness market product marketing areas churn

Sentiment Mobile Post-disas- High Social analysis Urban disaster Crime ter refugee frequency unrest Other of public planning relief detection reunification surveys prediction campaigns targeting

High Medium Low Pilot identified

Source: Naef et al . (2014), quoted in the “Measuring the Information Society 2014” report, ITU .

Harnessing the Internet of Things for Global Development 28 As for chronic conditions, people with type 1 diabetes Cellular networks are being used to transmit this data may soon be able to set their insulin doses by due to their wide availability and low costs . For example, smartphone . Researchers are testing a “bionic pancreas” in India, monthly cellular connectivity for connected pump that is inserted under the patient’s skin . When temperature monitor devices are in the range of around paired with an app and a small chip, this device is a dollar a month or less per fridge . The most expensive capable of tracking blood sugar levels and adjusting the component is often the cell radio, while conversely, amount of insulin and glucagon (another hormone that the temperature sensors are relatively cheap . In some controls blood sugar) on its own . A key study is slated cases, advantageous mobile cellular terms can be for 2016 and the researchers, who are based at Boston negotiated, resulting in reduced operating costs . The University and Massachusetts General Hospital, plan to wireless sensor device sends temperature data to submit the device for US FDA approval in 2017 7. Nexleaf servers and alert messages are sent to vaccine handlers and managers, if vaccine doses are in danger of In developing countries, one innovative way sensors spoiling . A weekly and monthly summary of refrigerators’ are being used is to monitor the ‘cold chain’ delivery temperatures is also sent via the website, so managers of vaccines, particularly to remote and rural areas . can understand how well equipment is functioning . Nearly one-fifth of children in the developing world go unvaccinated each year . One of the contributing These data can be used at the local and district level, factors to this problem and a large hurdle for healthcare as well as being aggregated up to or by the Ministry providers is vaccine spoilage, as many vaccines need of Health at the national level to determine how to to be stored at temperatures between 2 and 8 degrees allocate limited maintenance and equipment resources, Celsius . With over 200,000 vaccine refrigerators in use and where vaccine doses can be safely delivered (i e. . in the developing world alone, most of them in harsh and to ensure a batch of vaccines is not dispatched to a remote environments, keeping the cold chain up and broken refridgerator) . Data analytics are available on the running is a major challenge . Nexleaf Dashboard to help understand managers where problems are arising in power supply, and how to resolve A number of projects have focused on monitoring the these challenges . Another example of these types of cold chain, and one organization in particular, Nexleaf, interventions is the SmartConnect project, which focused has been working on remote sensing products for use on developing a “communication appliance” to improve 8 in difficult environments . In many countries, refrigerator the reliability and performance of the “cold chain .”9 temperatures are tracked and recorded by hand, often with delays in collecting the records . Now, using IoT IoT technologies are also being used to address technologies, Nexleaf’s ColdTrace system monitors and immediate challenges in humanitarian response, such records the refrigerator temperatures and send out SMS as the Ebola outbreak in West Africa . The United States alert messages whenever there is a problem (i e. ., when Agency for International Development (USAID) has temperatures rise above a predetermined threshold) . supported and employed IoT solutions via connected Nexleaf has developed a mobile-enabled temperature wearable technologies . Sensor Technology and Analytics sensor device that uploads data to a cloud-based serve to Monitor, Predict, and Protect Ebola Patients (or in near real-time via GPRS or SMS . The server sends STAMP2 for short) has been tested on Ebola patients regular messages to designated recipients regarding in the United States and is being scaled up to meet the current temperatures, and also warnings when the needs of government agencies such as USAID for its temperature exceeds critical thresholds . Ebola treatment strategy in Liberia .

STAMP2 collects patient data, including ECG, heart rate,

oxygen saturation, body temperature, respiratory rate, and position . These data are sent to a central server or 7 www webmd. com/news/breaking-news/future-of-health/default. . htm#wireless-medicine-toc/wireless-medicine

8 Interview with Nexleaf on March 23, 2015 . http://www .nexleaf org. 9 http://homes cs. washington. edu/~eorourke/papers/smart_connect_nsdr. .pdf

29 Harnessing the Internet of Things for Global Development platform so they can be monitored and analyzed over Water and Sanitation a long period of time and alert physicians of abnormal Nearly one billion people in the world lack access to safe changes in a patient’s behavior or health . The STAMP2 drinking water, while some two billion have inadequate sensor will be deployed in Ebola-stricken areas using access to sanitation facilities . Currently, some of the most a connected health patch, or “Smart Band-Aid .” The extensive uses of the IoT in developing countries are in fully equipped sensor-enabled band-aid is estimated projects where the objectives include the improvement of to cost approximately US$100 with a maximum battery clean water delivery and/or sanitation . life of ten days, making it ideal for use in field-based Ebola treatment centers . Deploying solutions such as In Bangladesh, a biosensor network of 48 manual the STAMP2 sensor can improve the Ebola response arsenic sensors is being used to monitor water quality 12. initiative at-large by decreasing emergency response In Jiangsu, China, water supply is being monitored by time in critical areas and enabling emergency responders adding IoT sensor devices at key points to register to detect Ebola patients earlier and monitor them more data on water usage and flow rates 13. In India, Sarvajal efficiently 10. Other examples include the use of remote has developed low-cost reverse osmosis technology diagnostics systems that allow for community health to provide clean water in rural areas, as well as smart workers to take measurements from patients and transmit meters to remotely monitor the quality and quantity of the data to doctors or specialists elsewhere .11 water . Additionally, a (WSN) is being used in the country to improve water management 10 http://mobihealthnews com/40564/scripps-wins-usaid-grant-to-monitor-. ebola-patients-with-medical-wearables/ and http://www .biospectrumasia com/. biospectrum/news/220757/usaid-unveils-wearable-technologies-tackle-ebola 12 http://users .ictp .it/~mzennaro/WSN4D .pdf 11 http://www .ictworks org/2015/07/31/5-mhealth-innovations-using-mobile-. phone-extensions-and-wearables 13 www chinamobileltd. com/en/ir/reports/ar2010/sd2010. .pdf

Figure 11: Monitoring the Movements of People during the Ebola Outbreak

The figure shows regional travel patterns in West Africa based on mobile phone data and demonstrates how anonymized call detail records (CDRs) can inform healthcare response activity in the event of an outbreak of infectious disease, such as Ebola .

Source: http://eprints .soton .ac .uk/370053/1/CommentaryEbolaOutbreak .pdf

Harnessing the Internet of Things for Global Development 30 Using a basic accelerometer, similar to that found in a mobile phone, to capture movement of the pump handle, the team developed a robust transmitter which fits into the handle of existing hand pumps, opting for “field readiness” over technological sophistication (e g. . using SMS rather than 3G to transmit data), as their need for more robust, reliable data outweighed that for data with advanced functionality . The team opted for non-rechargeable batteries, with a replacement lifetime longer than the pump maintenance cycle . An advantageous mobile package and efficient use of the SMS message format has resulted in relatively affordable data costs .

From the initial question of whether the pump was working or not, the team quickly realized that they could remotely monitor a host of data relating to water supply and demand, including hourly flow rates, usage data, hand pump performance, seasonality, and peak periods for demand . Based on these data, a more evidence-based approach to Photo credit: Tim Foster, Rob Hope, Johanna Koehler and policy decision-making has been achieved . It is no longer Patrick Thomson . Sensor enabled village hand water pump in Kyuso District, Kenya . the village that “shouts loudest” or that has the best social connections that ultimately receives service . The project in poor and semi-arid areas . The wireless COMMON- has been able to provide evidence about where and when Sense Net has been deployed over a small area of two the greatest needs for water are experienced . acres to measure temperature, humidity, ambient light, and barometric pressure in rural Karnataka . Soil moisture The team emphasizes that the project has led to a shift has been measured with a special probe since April in mindset . Although the Kenyan water regulator had 2005 14. Data from the sensors are visualized on the an existing mandate for regulating water supply in rural project’s website for real-time monitoring . areas, it was not able to engage effectively, as it lacked concrete data about the situation . Now that more data In Africa, water service reliability is closely correlated are forthcoming, the water regulator is better able to with extreme poverty and water insecurity in rural areas . manage resources . Performance-related pay has also Around one million hand pumps supply water to over been introduced for maintenance staff, who now know 200 million rural water users across the continent, yet that pumps are being monitored remotely, and who are as many as one third of all hand pumps are thought to generally more keenly engaged and responsive in repairing be broken at any given time, with 30-70% of pumps pumps . breaking within two years . The Smith School Water Although project staff were initially worried about Programme at Oxford University launched a 12-month vandalization of the equipment, once people saw that the ‘smart hand pump’ trial in Kyuso, Kenya, in 2013 in an technology worked and that it contributed to the pump attempt to resolve problems related to broken water being repaired more quickly, local social structures of pumps and to test a new maintenance model for respect and trust extended to include and protect the universal and reliable water services .15 transmitter . Due to these factors, over its lifetime, this project has helped achieve: 14 http://itidjournal org/itid/article/viewFile/244/114. http://ieeexplore ieee. org/stamp/stamp. .jsp?tp=&arnumber=4085509 ●● a ten-fold reduction in hand pump downtime (measured 15 www .smithschoolox . .ac .uk/library/reports/SSEE_Rights%20to%20Results_ by the number of non-functional days); FINAL_March2014 pdf.

31 Harnessing the Internet of Things for Global Development ●● a shift to 98% of hand-pumps functioning (up from For order/manufacture volumes of several hundred 67%); thousands of sensors at a time, price reductions can begin to be realized, which will help boost scale . ●● a more fair and flexible payment model contingent upon service delivery; and In Rwanda, sensors may add 10% to the cost of a hand

●● new and objective metrics to guide water service pump but may enable uptime to increase by some regulatory reform . 80-90%, significantly reducing the cost per unit of water per 10,000 litres delivered . Sensors are subjected MoMo is a similar, but fully mobile, device with a sensor to very harsh conditions, typically resulting in some that collects data to track infrastructure and improve 12-18 months of battery life . In Kenya and Rwanda, the accountability in the developing world . The device regulatory challenges have proven relatively limited so identifies where pumps are broken and alerts repair teams far, and the regulations for large-scale connectivity are to fix them . Data from MoMos can also help communities nascent . monitor the effects of infrastructure projects and inform 16 future investments . Water flow sensors are also being used in sanitation projects focused behavior change . For example, in one In Rwanda, SWEETSense uses sensor technology water, sanitation, and hygiene programme in Indonesia, developed by Portland State University’s SWEETLab flow sensors were combined with motion detectors to (Sustainable Water, Energy & Environmental Technologies measure the impact of behavior change training, with Laboratory) to monitor pump performance and water flow, the aim of increasing hygienic actions (i e. ., washing notifying technicians via SMS and emails 17. SWEETSense hands after latrine use) . The study found community technology uses sensors to provide continuous data participants were washing hands after latrine usage, but on usage and performance of programmes in water, that the survey responses (those who indicated they sanitation, household energy and rural infrastructure wash their hands after using the latrine) were significantly programmes with diverse partners including USAID, the higher than the data captured by the motion sensors, UK’s DFID, Mercy Corps, the Lemelson Foundation, Gates suggesting that over-reporting had occurred in the verbal Foundation and DelAgua Health in India, Nepal, Indonesia, and self-reported surveys 18. Similar techniques can Kenya, Rwanda and Haiti . Indeed, SWEETSense be used to remind staff of basic hygiene techniques in technology has been used for: homes, clinics, and hospitals .

●● Water pumps in Kenya; Agriculture ●● Cooking stoves in India; In agriculture, IoT technologies can be used to increase, ●● Latrine monitoring in Bangladesh; and protect, and optimize crop production, as well as improve ●● Water filters for hand-washing stations in Indonesia . the storage and distribution of food . Growth in agricultural productivity over the last fifty years has been much These sensors use Wi-Fi or cellular via GSM using a slower in developing regions of the world, in part due local SIM card (in East Africa, Airtel or MTN or Safaricom to large capital costs 19. Similarly, gathering and utilizing offer data plans for around US$6-10 a month to transmit local weather data, a critical aspect of farming, remains data) . The data is then integrated into SWEETData™, a major challenge in developing regions due to limited an Internet database monitoring summary statistics on coverage . Traditional weather monitoring equipment is performance and usage to front-end users . Sensors large and capital-intensive, but the IoT is now allowing currently cost more than US$100, but this should fall relatively quickly, as demand increases .

18 Evan Thomas; Kay Mattson 2014 . Instrumented Monitoring with Traditional 16 Interview with WellDone org/. Momo on March 24, 2015 . Public Health Evaluation Methods . Available at: http://www .mercycorps org/sites/. https://www welldone. org/. default/files/Instrumented%20Monitoring%20Indonesia .pdf

17 http://newsroom cisco. com/feature/1556125/Sen. - 19 http://www economist. com/news/middle-east-and-africa/21665005-small-. sors-Change-Lives-in-Developing-Countries farmers-africa-need-produce-more-happily-easier-it

Harnessing the Internet of Things for Global Development 32 for micro-weather stations to be deployed and utilized about 25 million water pumps are in use for farm for a range of activities, including the dissemination of irrigation . Many of these pumps have to be manually information to farmers on nutrient requirements, the operated, based on rainwater conditions, electricity prediction of weather patterns, and the provision of availability and crop needs . For the average small-scale inputs into localized crop insurance schemes . farmer, the variability of these factors on a day-to-day basis adds extra burdens in terms of time, labor and For example, Syngenta’s Kilimo Salama (“Safe fuel costs . In many cases, farmers need to travel long Farming”) project is a connected weather station that distances through difficult conditions to access their monitors agricultural events and facilitates linkages pumps from their households . with insurance firms . The aim is to mitigate the risks associated with adverse weather, thereby providing a The Nano Ganesh unit works by attaching to the irrigation much-needed safety net for farmers while promoting pump, and serving as an actuator which can turn the agricultural investment and improved livelihoods . pump on and off via basic commands from a farmer’s Safaricom’s M-Pesa mobile banking system assists simple feature phone (2G mobile telephones) . The Kilimo Salama in keeping index insurance premiums farmer is also able to check the availability of electricity more affordable, helping transform smallholder farmers at the pump, as well as the availability of water near the into a commercially viable market segment for insurance pump (with an additional water sensor) . By August 2014, firms 20. around twenty thousand farmers in India had benefitted from Nano Ganesh . Various types of micro-weather stations capture a range of data such as air and soil temperatures (oC and oF), China is making great strides in applying IoT technologies air and soil moisture levels (%), solar radiation (W/m2), to improve agricultural production . Various information- wind direction, wind speed (m/s), atmospheric pressure based applications have been developed, including (hPa), amount of rainfall (mm), soil electrical conductivity greenhouse remote monitoring, automatic drip irrigation, (EC, 0-23dS/m), and visual appearance (image capture) . and milk source safety information management to At the end of each growing season, weather statistics enhance agricultural production . In Xinjiang, the “mobile collected from solar-powered weather stations are Internet of Things for Agriculture” project uses wireless automatically compared with an index of historical monitoring of agricultural greenhouses . Wireless water- weather data . Rainfall measurements are factored into saving drip irrigation has also been used since 2011 to specialized agronomic models to determine the impact monitor water quality and to save water in fresh water and likely loss that farmers experience . Insurance aquaculture 22. payouts are then calculated and sent to the insured farmers via automated mobile payments . This mechanism Additional examples include the use of RFID tags for has effectively automated and simplified the claims monitoring livestock, which allows for more personalized process, cultivating a financially supportive environment care for individual animals . In tea plantations in Sri for individual farmers and encouraging agricultural Lanka and Rwanda, WSNs are being utilized to monitor production at all levels . soil moisture, as well as carbon, nitrogen, potassium, calcium, magnesium, and pH levels . The sensors and In India, Nano Ganesh is a low-cost solution to provide connectivity modules are powered through solar panels, small-shareholder farmers with a tool that can remotely and the data are transmitted wirelessly 23. control their micro irrigation pumps .21 Across the country,

20 “The Broadband Effect: Enhancing Market-based Solutions for the Base 22 http://www chinamobileltd. com/en/ir/reports/ar2010/sd2010. .pdf of the Pyramid ”. Available at: https://publications .iadb org/bitstream/han. - http://newsroom .hwtrek com/?p=626. dle/11319/6642/Opportunities_for_the_Majority_Report_The_Broadband_Ef- fect .pdf .pdf?sequence=1 23 Minuri Rajapaksa “IoT for Productive Tea Plantation .” Available at: http://wireless .ictp .it/school_2015/presentations/CaseStudies/IoTforTeaPlanta- 21 Food and Agriculture Organization of the United Nations, “Success Stories tion-SriLanka-Minuri .pdf on Information and Communication Technologies for Agriculture and Rural http://wireless .ictp .it/rwanda_2015/ Development “. Available at: http://www .fao org/3/a-i4622e. .pdf http://www .ictp .it/about-ictp/media-centre/news/2015/6/teatime-with-iot .aspx

33 Harnessing the Internet of Things for Global Development Resiliency, Climate Change, economic exposure to risk 26. New sensor data also and Pollution Mitigation include unmanned aerial vehicles (“drones”) and spatially referenced (geo-referenced) video . Following the 2004 Indian Ocean tsunami that devastated coastal areas in India, Sri Lanka, Thailand and Indonesia, Geo-referenced video has been used to quickly identify the international community united to establish the sites of standing sewage and water to aid in cholera Indian Ocean Tsunami Warning System whereby kinetic risk mapping in Haiti 27 and the vulnerability of homes sensors (measuring waves and water flow) placed on in Los Angeles, to wildfire 28. Drones can provide very the ocean floor communicate data on potential tsunamis high-resolution (VHR) satellite imagery in 2-D and 3-D, to disk buoys floating on the ocean surface via acoustic which can be useful in mapping complex urban riverine telemetry . The buoys then upload the information to topography, and which has been used in Haiti for flood government authorities via satellite connectivity 24. The ITU has launched a project to detect earthquakes and seismic events via a network of sensors hosted on submarine cables .

Other applications designed to promote resiliency include a Red Cross project to explore the widespread installation of connected alarm systems across high density urban slums to quickly notify residents of fast-moving fires . Fires can move quickly in informal settlements and slum areas, given that homes are close in proximity . Faulty wiring and indoor open hearths, when combined with the density of these settlements, make combating fires quite difficult, and, unfortunately, more likely to start . The Red Cross is exploring the development of low-cost, solar- powered sensors networked together to quickly detect and relay to authorities when fires emerge . The network sounds alarms, communicates to threatened residents (via SMS and other modalities), and its connected sensors identify via GPS where the fire has started, notifying authorities of the location where fire mitigation efforts should be targeted 25. Currently, the intervention is Photo credit: Marco Zennaro . Calibration of a sensor node being tested in Nairobi and Cape Town, with participation measuring weather parameters (temperatures, pressure, by two thousand households . humidity, light) in Nairobi, Kenya .

Many analyses are now trying to model risks and vulnerabilities in the wake of climate change . For 26 Ceola, S ., Laio, F ., & Montanari, A . (2014) . Satellite night-time lights reveal increasing human exposure to floods worldwide, 7184–7190 . example, climate models and disaster risk models doi:10 1002/2014GL061859. . Received; and Christenson, E ., Elliott, M ., Banerjee, O ., Hamrick, L ., & Bartram, J . (2014) . Climate-related hazards: can now be combined with satellite imagery of human a method for global assessment of urban and rural population exposure to settlement (such as night-time lights) to estimate cyclones, droughts, and floods . International Journal of Environmental Research and Public Health, 11(2), 2169–92 . doi:10 .3390/ijerph110202169

27 Blackburn, J . K ,. Diamond, U ., Kracalik, I . T ., Widmer, J ., Brown, W ., Mor- rissey, B . D ., … Morris, J . G . (2014) . Household-Level Spatiotemporal Patterns of Incidence of Cholera, Haiti, 2011 . Emerging Infectious Diseases, 20(9), 1516–1520 . 24 http://www .kophuket com/phuket/homemenu/tsunami. .html and https:// en wikipedia. org/wiki/Indian_Ocean_Tsunami_Warning_System. 28 Burkett, B ., & Curtis, A . (2011) . Classifying Wildfire Risk at the Building Scale in the Wildland-Urban Interface: Applying Spatial Video Approach- 25 https://drive google. com/file/d/0B1vf6TLGIC0yZk9UU2t2UGFmNlE/ . es to Los Angeles County . Risk, Hazards & Crisis in Public Policy, 2(4), 1–20 . view?usp=sharing doi:10 .2202/1944-4079 .1093

Harnessing the Internet of Things for Global Development 34 modelling, as well as in Nepal after the destructive 2015 Additionally, privacy issues still need to be addressed, earthquake .29 such as the right level of anonymity (and aggregation, such as from the individual to a group level) for records to New datasets can help in understanding vulnerability ensure an appropriate balance is struck between ensuring and mobility, and data to estimate mobility patterns can individual privacy, while preserving the value of the data be gleaned for example from geo-located tweets . One for social aims (including crisis response and policy researcher analyzed New York City tweeters before, planning) 34. during, and after Superstorm Sandy to show that pre- disaster mobility patterns can indicate the potential range Urban air pollution is a major problem in many developing of mobility during a disaster 30. Other indicators of mobility country cities . The World Health Organization (WHO) include transit data by bikes,31 buses and subways being suggests that polluted air contributes to one in eight made available by hundreds of municipalities 32. Transit deaths worldwide, as dirty air causes lung damage, heart data can monitor population flux at different times of day, disease, strokes and cancer . The WHO also estimates and provides just one example of open data which cities that indoor air pollution in homes in Africa contributed to are releasing that could be valuable for risk assessment . nearly 600,000 deaths in 2012 35. To measure the extent of the problem, air quality sensors are being deployed in Population movement analyses based on Call Detail a range of cities to track levels and changes in pollutants . Records (CDRs) from mobile network operators have One such project, “Fresh Air in Benin“ is focused on also been used for planning malaria elimination strategies developing a network of air quality sensors to capture and, more recently, to monitor population movements in and send data every 20 minutes via GSM connectivity 36. the 2014 Ebola outbreak in Guinea, Sierra Leone, Liberia, Nigeria and Senegal . Despite some initial challenges, Water flow sensors are also being used to help collect epidemiological models of the spatial spread of Ebola were hydrological data in developing countries where local developed to model the spread of the virus, and predict its data on river flow and levels may not be regularly possible development (Figure 11, page 30) . These models collected . These sensors can also provide early warning can help assess the likely routes of infected individuals of floods . For example, the Hidrosónico is a water stream between populations, predict possible new outbreaks and gauge that uses a sonar range sensor to measure the help focus the delivery of eventual vaccines . However, distance to water surface level . The module sends the challenges remain in terms of establishing the processes readings on a regular basis to recipients (via SMS or by which such data can be shared and released in a timely email) or to a cloud application . The unit is also equipped fashion 33. with a rain gauge to monitor precipitation, and is currently deployed in Honduras 37.

29 Big Data in the Disaster Cycle: Overview of use of big data and satellite im- aging in monitoring risk and impact of disasters; Tellman, Schwarz, Burns, Adams

30 Wang, Q ., & Taylor, J . E . (2014) . Quantifying Human Mobility Perturbation and Resilience in Natural Disasters . arXiv org,. physics .so(11), 1987 . doi:10 .1371/ journal .pone .0112608 34 http://www .unglobalpulseorg/projects/mobile-data-privacy .

31 Zaltz Austwick, M ., O’Brien, O ., Strano, E ., & Viana, M . (2013) . "The struc- 35 http://www .nytimes com/2014/04/17/business/energy-environment/mea. - ture of spatial networks and communities in bicycle sharing systems . PloS One, suring-africas-air-pollution .html 8(9), e74685 . doi:10 .1371/journal .pone .0074685 36 http://www worldbank. org/content/dam/Worldbank/Feature%20Story/Mak. - 32 https://code .google com/p/googletransitdatafeed/wiki/PublicFeeds. for a list ers4development .FV%20USAID .pdf

33 http://www economist. com/news/science-and-technology/21627557-mo. - 37 http://dai com/our-work/solutions/dai-maker-lab. and https://github com/. bile-phone-records-would-help-combat-ebola-epidemic-getting-look DAI-Maker-Lab/hidrosonico

35 Harnessing the Internet of Things for Global Development Natural Resource Management connected drones for surveillance as well as long-range wide area technologies to track wildlife and monitor Protecting land and the environment from pollution activity at the boundaries of game parks 40. and illegal logging, as well as protecting natural wildlife from poaching, are major natural resource challenges In Timor-Leste, the National Directorate of Fisheries and in developing countries . One example of a working Aquaculture (NDFA) is working in partnership with the prototype drone has been developed to help monitor FAO-Spain Regional Fisheries Livelihoods Programme wildlife in remote and mountainous areas in the United for South and Southeast Asia (RFLP) to introduce a Arab Emirates (UAE) . Camera traps take automatic programme using radio location beacons to protect local photos of animals, triggered by motion . However, waters from illegal fishing and to provide emergency reaching the camera traps is dangerous and costly and it response to fishermen in distress . The partnership is time-consuming to upload the photos, costing at least introduced low-cost personal location beacons (PLBs) Dh1 million a year to capture this data . To combat these as part of a community-based system for identifying challenges, a team from New York University in Abu illegal, unreported and unregulated fishing . When local fishermen identified illegal fishing boats in the Timor- Dhabi has developed a 2 .2kg drone that can fly for 1 .5 Leste waters, the location beacon is activated and data hours for up to 40 km, collecting different types of data is are sent to an Internet-enabled mapping platform used including images, salinity and atmospheric data in the by authorities to track and apprehend illegal fishing boats, mountainous Wadi Wurayah National Park 38. or initiate rescue services 41. In the camera trap, the transmission system uses Acoustic sensors have also been developed to monitor low-power XBee communications run on solar power, wild populations of seabirds . For example, Nexleaf has which turns on at regular 2 to 3 minute intervals to check worked with the US Fish & Wildlife Service and the whether the drone is approaching . The system also has a Coastal Conservation & Action Lab to deploy four sensors back-up transmission system via Wi-Fi in case of failure . on Tern Island from December 2011 to May 2012 42. The drone hosts the same XBee and Wi-Fi systems and The deployment consisted of four wildlife acoustic a microprocessor that is permanently turned on and monitors and two Wi-Fi radio repeaters powered by 20W powered by a litho-battery, which can be recharged from solar panels, and one gateway computer . The project solar power once the drone returns to base . The drones demonstrated the performance capabilities of satellite- are currently used to assess and evaluate the animal based WSNs for long-term monitoring of seabird population, rather than climate change per se . However, colonies . new applications could include monitoring water quality and salinity in coastal regions and desalination of the The United States Forest Service (USFS) has developed coast . a project to study the state of urban and rural forest ecosystems in real-time to keep them healthy and Illegal poaching of large wildlife is a major concern across more sustainable 43. Climate change will also be better sub-Saharan Africa . For example, 40,000 elephants understood, and lessons can be learned in order to take were killed illegally in 2014, primarily for their tusks, while advantage of the effects of climate change . USFS has demand for the horns of the black rhinoceros has lead to monitored wildlife across the US for decades, manually a 96% decline in the species’ population from 1970 to collecting information from air and soil temperature to 1992, with fewer than 20,000 animals remaining . Every eight hours, a rhinoceros is killed in southern Africa 39. 40 http://airshepherd org/. and http://gblogs cisco. com/uki/week-4-how-can-. To combat poaching operations, a wide number of technology-help-the-anti-poaching-activities/ projects are utilizing emerging IoT technologies such as 41 Food and Agriculture Organization of the United Nations, “Success Stories on Information and Communication Technologies for Agriculture and Rural Development”, available at: http://www .fao org/3/a-i4622e. .pdf 38 http://gulfnews com/news/uae/government/uae-drones-for-good-award-. wadi-drone-makes-conservationists-job-easier-1 .1452917 42 http://nexleaf org/project/tern-island.

39 http://airshepherd org/. 43 http://smartforests org/.

Harnessing the Internet of Things for Global Development 36 solar radiation . The IoT is helping to automate some of and communications model 46. Individuals purchase the data collection activities . the system at a discounted rate with the capital costs amortized over an initial purchase period . After installation Now, the USFS has developed a sensor-based system to in their dwelling, the customers are then able to utilize obtain real-time measurements and combine these with the electricity generated by the solar cells to power home traditional field studies and long-term records of patterns appliances . They must make regular payments (usually and processes to analyze environmental changes . Some via mobile money systems, e .g . M-Pesa or Airtel Money) sensors reach around trees to measure their growth, in order to continue using the device . M-Kopa is able others use motion-triggered webcams and sensors to to remotely monitor the amount of electricity captured/ capture pictures of wildlife 24/7 and track the presence stored and whether the device is working appropriately (or absence) of endangered species . Sensors detect through the device’s connectivity module . and deliver high-resolution data wirelessly or through a cellular network to a central web portal . Similarly, Across the developing world, wood and charcoal Rainforest Connection is using microphone sensors in burning cook stoves are used extensively to prepare smartphones to monitor for sounds associated with illegal meals and provide a source of heat inside homes . The logging (from chainsaws, trucks and motorcycles) 44. resulting indoor air pollution contributes to approximately 4 million deaths a year, out of the 3 billion people There are also other interesting developments in relation worldwide who utilize biomass to prepare their meals . to commercial crops, often a vital source of income in As a result, a number of initiatives are in place to help rural communities . Over the last two decades, the red lower income households use less polluting methods palm weevil has become a growing threat to palm trees for meal preparation and heating in the way of improved 45 in many parts of the world . Early detection is difficult, cookstove projects . For example, the US Government since many palm trees may not show visual evidence has supported a five-year “Global Alliance for Clean of infection until it is too late for trees to recover . A Cookstoves” which aims to achieve a goal of enabling prototype bioacoustic sensor has been developed to 100 million homes to adopt clean and efficient cooking detect sounds of larvae activity, with wireless reporting solutions by 2020 47. IoT sensors are playing a role in this to a control station to check the status and evolution of initiative by helping measure the black carbon emitted palm tree orchards . by cookstoves in real-time, as well as monitoring and evaluating projects to disseminate improved cookstoves . Energy Another interesting IoT application in the energy sector In one particular project in Sudan, the use of improved in developing countries has been the rapid adoption of cookstoves via sensor-enabled recording instruments off-grid solar panel systems that provide steady electrical was compared to traditional survey data (captured by power to low-income families . As Figure 8 (page 26) an enumerator) to determine whether the latter method highlights, electricity is used by only 58% of the urban accurately reflected usage captured by sensors . In this population and by only 12% of the rural population in case, the study found that survey participants were over- sub-Saharan Africa . Challenges of grid availability, cost of reporting daily cooking hours by 1 .2 hours on average, service and frequent service interruptions plague on-grid and daily cooking events by 1 3. events . The results electricity users across much of the developing world . suggest that survey-based methods of evaluation may M-Kopa in Kenya exemplifies this new technology, which be misstating the actual impact and usage of the newer 48 is comprised of photovoltaic cells and a battery system technologies .

46 www .gsmacom/mobilefordevelopment/wp-content/uploads/2013/01/Sus . - tainable-Energy-and-Water-Access-through-M2M-Connectivity .pdf

44 Interview with Dave Grenell, Rainforest Connection, on May 19, 2015 . 47 http://www .state .gov/r/pa/prs/ps/2015/09/247240 .htm https://rfcx org/. 48 “Comparing Cookstove Usage Measured with Sensors Versus Cell Phone- 45 www .mdpicom/1424-8220/13/2/1706/pdf . Based Surveys in Darfu, Sudan,” by Daniel Lawrence Wilson et al . Chapter 20 of “Technologies for Development: What is Essential .”

37 Harnessing the Internet of Things for Global Development IoT technologies can be used to produce more accurate data, especially when used in conjunction with subjective perceptions recorded via survey response .

Other Sectors Examples of IoT deployment impacting other facets of global development abound, and the opportunities to improve service delivery, and other aspects of development work, are limited only by human creativity and the resources available . Table 3 below highlights some of the various IoT interventions as they map to the Millennium Development Goals (MDGs) and the new Sustainable Development Goals (SDGs) as adopted by the United Nations .

Harnessing the Internet of Things for Global Development 38 Table 3 – Examples of IoT interventions mapped to the Millennium Development Goals (MDGs) and Sustainable Development Goals (SDGs)

Sector MDG SDG Examples

Sensor- and SMS-enabled village water pumps MDG 4: SDG 3: (Rwanda, Kenya); GSM-connected refrigeration for Child Health Ensure healthy lives and vaccine delivery in the ‘cold chain’ (Global); sensor- enabled ‘band aid’ to monitor Ebola patients’ ECG, MDG 5: promote well-being for all at all heart rate, oxygen saturation, body temperature, Health, Water Maternal health ages . respiratory rate and position, all remotely (West & Sanitation MDG 6: SDG 6: Africa); water stream gauge with sonar range sensor Combat HIV/ Ensure availability and to monitor river flow and depth (Honduras); water AIDS, malaria and sustainable management of water flow sensors and motion detectors in latrines to other diseases and sanitation for all . monitor efficacy of hygiene training and intervention (Indonesia) .

SDG 1: End poverty in all its forms Connected micro-weather stations improving everywhere . localized weather data and provision of crop failure SDG 8: insurance (Kenya); low-cost mobile-controlled micro Promote sustained, inclusive and irrigation pumps (India); soil-monitoring sensors MDG 1: Agriculture & sustainable economic growth, full used to improve tea plantation production (Sri End Poverty & Livelihoods and productive employment and Lanka, Rwanda); RFID-based food supply testing Hunger decent work for all . and tracking system (India) and RFID based livestock SDG 2: programmes for tracking, theft prevention and End hunger, achieve food security vaccination records (Botswana, Senegal and and improve nutrition, and Namibia) . promote sustainable agriculture .

SDG 4: Smart identity cards with biometric features for all MDG 2: Ensure inclusive and equitable public school students to improve service delivery Education Universal quality education and promote (Nigeria); biometric clocking device to improve Education lifelong learning opportunities for teacher attendance in real-time (South Africa) . all .

SDG 12: Ensure sustainable consumption and production patterns . SDG 13: Take urgent action to Radio-based cloud-connected devices to combat climate change and its identify and track the presence of illegal fishermen impacts . (Timor-Leste); air pollution sensors to monitor urban outdoor air pollution (Benin); acoustic sensors to SDG 14: monitor sea bird populations (global); sensors and Environment & MDG 7: Conserve and sustainably use connectivity to protect game park perimeters and Conservation Environment the oceans, seas and marine track animals (Africa); connected unmanned aerial resources for sustainable vehicles monitor national parks and connecting development . images from camera traps (UAE); acoustic sensors SDG 15: Protect, restore in tropical rainforests ‘listening’ for illegal logging and promote sustainable use (Indonesia) . of terrestrial ecosystems, sustainably manage forests, combat desertification, halt and reverse land degradation, and halt biodiversity loss .

39 Harnessing the Internet of Things for Global Development Table 3 – Examples of IoT interventions mapped to the Millenium Development Goals (MDGs) and Sustainable Development Goals (SDGs) (continued)

Sector MDG SDG Examples

SDG 7: Ensure access to affordable, reliable, sustainable and modern Networked fire/smoke alarms in high-density urban slums/ energy for all . informal settlements (Kenya, South Africa); Connected SDG 9: buoys as part of the tsunami monitoring system (Indian Resiliency, Build resilient Ocean); off-grid micro solar electricity systems for Infrastructure infrastructure, promote electricity for lower-income households (east Africa, India); and Energy inclusive and sustainable connected black carbon- and use sensors to monitor cook industrialization, and foster stoves (Sudan); sensor-connected matatus (mini-buses) innovation . tracking velocity, acceleration, and braking to curb danger- SDG 11: ous operation of public transportation (Kenya) . Make cities and human settlements inclusive, safe, resilient and sustainable .

SDG 10: Reduce inequality within and among countries . SDG 16: Governance & Promote peaceful and Retinal scans used for ATMs providing secure biometric Human Rights inclusive societies for cash assistance to displaced refugees (Jordan) . sustainable development, provide access to justice for all and build effective, accountable and inclusive institutions at all levels .

SDG 5: Achieve gender equality MDG 3: and empower all women Gender Equality and girls . Cross-Cutting MDG 8: SDG 17: Partnership Strengthen the means of implementation and revitalize the global partnership for sustainable development .

Harnessing the Internet of Things for Global Development 40 Challenges to the deployment, scale and impact of the IoT in developing countries

41 Harnessing the Internet of Things for Global Development Challenges of the Internet of Things can help the development of the IoT . However, outdated or poorly designed frameworks can prove a hindrance Despite all the exciting possibilities brought about by and obstacle to the further growth of the IoT . While many the IoT and Big Data, signifi cant challenges persist . The parts of daily life become more connected, some remain same infrastructure that enables people to create, store woefully underconnected . Conversely, other elements and share information may also jeopardize their privacy of an individual’s daily life may be overwhelmed as the and security . These same techniques can be used for explosion of new devices will require new infrastructure large-scale and targeted surveillance . Abuse of these and technologies . techniques could turn the ‘Information Society’ into the ‘Surveillance Society’ , as identity management systems Technological and human capabilities are often improve without parallel emphasis on anonymity and insuffi cient in developing countries . Financial support may ownership of personal data . be lacking . There are often not enough technically literate people with IT skills in local areas who are capable of Society’s most advanced systems and infrastructures are implementing the use of sensors or other devices into now so complex that some of them are becoming hard their daily lives . Figure 12 summarizes some of the to manage eff ectively . Where they are designed wisely emerging challenges in relation to the IoT and data . and used eff ectively, policy and regulatory frameworks

Figure 12: Summary of Emerging Challenges in relation to the IoT

Data Reliability Localization

Standards Access to data/ Scaling Open data Interoperability Power Legacy Security Regulatory Technical Connectivity Models Policy Privacy Cost IPR Spectrum & Bandwidth Cross - Capacity Constraints border tra­c

IPv6 Governance

Source: Pepper, R . & Garrity, J . (2014) The Internet of Everything: How the Network Unleashes the Benefi ts of Big Data . Global IT Report 2014 . WEF . http://www3 .weforum .org/docs/GITR/2014/GITR_Chapter1 .2_2014 .pdf

Harnessing the Internet of Things for Global Development 42 Technical Challenges mysterious ‘black boxes’ without back-up strong structures in place to deal with breakdowns . Reliability is a concern with regard to the durability of devices to withstand external conditions . Sensors, too, Further challenges may arise from human behavior as a need to be calibrated to ensure proper measurements . limiting factor, with reluctance to adopt new technology In terms of scalability, the way in which resources are also a possible concern . People are often resistant and scaled to match growth in the IoT may matter . Data reluctant to modify their behavior to fit with systems, and centres, for example, are constantly being redesigned prefer that systems adapt to meet their needs . Deloitte in terms of electrical power, cooling resources, and cites the example of one electrical utility company which space design to advance current capabilities . However, installed smart meters in millions of homes in North the connectivity requirements of billions, as opposed America, expecting that consumers would consult online to millions, of connected objects impose very different dashboards of monthly usage, and modify behavior demands on data centres . As the IoT scales up and to save money and energy, benefit the environment . expands from billions into tens of billions of connected Three years after the meters were deployed, only 6% of devices, IP networks have to be able to manage the huge households had viewed the dashboard at all, while fewer 3 scale of device connectivity . than 2% had consulted the dashboard more than once . Indeed, according to some projects, human behavior may Power requirements vary greatly, with higher bandwidth prove a more significant barrier to adoption than some devices requiring much more power . Connectivity technical challenges . challenges were discussed earlier, and include limited data network coverage . According to Laura Hosman In the short-term, the transition to IPv6 has proved of Inveneo, the top five hardware challenges in the challenging for some countries and some organizations to application of ICTs in development are: electricity/ date . However, practically speaking, IPv6 may even prove power/energy; cost; environment; connectivity; and a limiting factor in some M2M deployments, requiring maintenance and support 1. The costs associated with all partners in a project to have made the transition . For the sensors, connectivity modules and the connectivity example, partner universities first need to transition to IPv6 service can still prove prohibitive for many interventions before they can implement a WSN project . Government (such as for individual small shareholding farmers) . can help play a leadership role in the transition to IPv6, and Organizations are starting to explore shared models of help aggregate and/or stimulate demand . The migration sensor module ownership such as community ownership, from IPv4 to IPv6 should help resolve Internet address or ‘sensors as a service’ .2 issues in the long-term, but the short-term challenge of providing adequate address space for billions of objects Inadequate human capacity may prove a major issue in will persist . some locations . Small-scale organizations may not be trained properly to use the technology . There may also Policy Considerations be underlying issues that inhibit training . For instance, if Data localization regulations and limits on cross border 80% of the target population is illiterate, is SMS text really traffic may impede the ability of managers to send data the best form of communication? There may also be an to cloud-based servers where data may be analyzed . inadequate number of trained people or technicians to While Open Data policies are increasing adopted, there respond, once a system signals a problem . If it is difficult are examples of governments clamping down on access to fix manual pumps on-the-ground, it may be difficult to data collected and generated by sensors 4. to find the resources to fix more complicated, seemingly The IoT has regulatory implications across a set of areas,

1 http://www .inveneo org/wp-content/uploads/2014/07/FINALTop-ICT-. Hardware-Challenges-White-Paper .pdf 3 Deloitte Tech, Media & Telecoms (TMT) Predictions 2015 .

2 http://www .slatecom/articles/technology/future_tense/2015/06/community_ . 4 www washingtonpost. com/blogs/monkey-cage/wp/2015/04/08/five-charts-. drones_helps_indonesia_s_dayaks_protect_their_land .html that-may-soon-be-illegal-in-tanzania/

43 Harnessing the Internet of Things for Global Development including licensing, spectrum management, standards, to interoperability, and can achieve high levels competition, security and privacy — only some of which of interoperability 8. are the familiar territory of telecom regulators, compared with other areas where different regulators may typically Privacy and security are often cited as two of the most take the lead . The ‘full’ IoT or Internet of Everything significant (and closely related) issues in large-scale (IoE) is likely to require more ‘joined up’ regulation, IoT deployment . Privacy, security and anonymity are with telecom/ICT regulators working more closely with all related, but separate, concepts . Privacy (related to their counterparts in data protection and competition, confidentiality) is the ability to define the intended target but also with emergency services, health and highway audience for data . Anonymity is the quality or state authorities, as legacy regulatory models (e .g . power utility of being unknown to most people . A secure system regulations) may prove inadequate to deal with emerging is a system free from weakness or vulnerability . For technologies (e .g . ‘smart grid’ technologies) . example, electronic health files and car license plates are private (and confidential), but not anonymous . Similarly, broader governance issues may impede the A breach of security may or may not result in a loss adoption of the IoT, such as in the case of the slow of privacy, depending on the data downloaded and adoption of connected thermometers to protect the how it is subsequently used . While ICTs do provide vaccine cold chain due to challenges in certification greater opportunities for communication and income approval for new technologies in the World Health related activities for lower-income populations, careful Organization’s Pre-Qualified Systems (PQS) . consideration is needed of the risks associated (loss of privacy, etc ). . Overlapping Issues Without adequate security, intruders can break into IoT The proliferation of, and growth in, new IoT technologies systems and networks, accessing potentially sensitive are built mainly on interoperability . In order for a car, personal information about users, and using vulnerable airplane, parking meter or pill bottle to send and receive devices to attack local networks and devices, thereby important data, it needs to be able to connect to breaching their privacy . This is a particular issue when other systems and networks seamlessly and securely . devices are used in private spaces, such as in individuals’ Interoperability is the ability to transfer useful data and homes . Operators of IoT systems, as well as others with other information across systems, applications, or authorized access to the data produced, are likely to be components in four broad layers — technological; data; in a position to “collect, analyze, and act on data from human; and institutional 5. According to one source, within previously private spaces” 9. there are at least 115 different protocols used by IoT 6 devices to connect to the cloud today . A recent paper Privacy concerns may still arise, even where systems are by McKinsey notes that “interoperability is required secure and functioning as intended . For example, does to unlock more than US$4 trillion per year in potential sponsorship of projects by third parties entitle them to economic impact for IoT use in 2025, out of a total access the resulting data? Third parties almost invariably impact of US$11 .1 trillion . On average, interoperability is respond that the data are high-level and aggregated, necessary to create 40% of the potential value that can making it anonymous, or that they are using the data for 7 be generated by the IoT in various settings .” beneficial purposes that outweigh any loss in privacy .

Although various approaches exist to help promote Another issue related to privacy and anonymity is the interoperability, standards are one collaborative approach address space of connected devices . Identifiers used in one network need to be understandable and/or usable 5 GSR Interoperability Discussion Paper by Urs Gasser, Berkman Center . (i e. . interoperable) in another network . In the Internet of 6 NetHope SDG ICT Playbook . Available at: http://solutionscenter .nethope org/. assets/collaterals/NetHope_SDG_ICT_Playbook_Final .pdf 8 GSR Interoperability Discussion Paper by Urs Gasser, Berkman Center . 7 Manyika, James, et . al ., The Internet of Things: Mapping the Value beyond the Hype . McKinsey Global Institute, June 2015 . p . 2 . 9 GSR Discussion Paper on IoT by Prof . Ian Brown, Oxford Internet Institute . .

Harnessing the Internet of Things for Global Development 44 Things, consumers will likely want to use different objects It could also be very difficult to renegotiate the terms of across multiple kinds of heterogeneous networks, which the mobile contract, or swap service temporarily to take will need the identities of things to be “federated” or advantage of better service from a different provider, capable of being translated accurately and recognized raising concerns of anti-competitive behaviour . by different networks . The IoT will contain billions of objects that must be uniquely identified, a challenge for Spectrum and bandwidth requirements may impede which there is currently so far no internationally-agreed the adoption of IoT devices and services . According solution, although Internet Protocol version 6 (IPv6) may to Cisco’s 2015 Visual Networking Index study, over eventually become the default solution . 10 billion new devices will come online between 2014 and 2019, and total global IP traffic is growing at 23% On a more practical level, lock-in to mobile service Compound Annual Growth Rate (CAGR) .10 While the providers is a real likelihood for many projects . Firms sheer diversity of IoT devices will result in a wide range operating large networks of IoT devices via mobile of network requirements, the aggregate impact will lead telephony networks using a fixed SIM in each device to increasing demands for wireless spectrum to support may not find it easy to switch network at the end of a wireless data transmission . A flexible and sufficient contract, or where there is device roaming in different spectrum bandwidth regime will be necessary to ensure network areas . innovation and adoption are not stifled .

10 Cisco Visual Networking Index 2015, at http://www cisco. com/c/en/us/solu. - tions/service-provider/visual-networking-index-vni/index .html

Photo credit: Marco Zennaro . Engineer with a GSM-enabled soil moisture sensor node in a tea factory plantation in Rwanda .

45 Harnessing the Internet of Things for Global Development Recommendations

Harnessing the Internet of Things for Global Development 46 A number of enabling policies can be put in place to the IoT . Efficient and effective spectrum management facilitate the deployment of the Internet of Things, as well is therefore key for ensuring efficient connectivity . as Big Data . As explained previously, policies to support In December 2015, Australian authorities freed up rapid and effective adoption of the IoT need to involve additional spectrum bands dedicated to the use of a range of stakeholders to help promote successful IoT and M2M . deployments . At the national level, policies can be put in 3 . Promote and support a broad, vibrant ecosystem place to support and facilitate the fast development of for IoT, including support for tech start-ups and IoT, as well as eliminate the barriers and challenges to incubators . This includes promoting policies to be overcome (described in the last section) . Some key facilitate innovation and development and eliminate aspects to be considered include: policies that restrict or prevent innovation (such as 1 . Create a policy framework and master plan restrictions to the free movement of data or the addressing the IoT – as the IoT is likely to become a ability to trade in digital services) . huge sector on its own right in a number of countries, 4 . Promote Data Centres — Depending on the countries wishing to take advantage of its benefit centralized or decentralized nature of the IoT should develop a master plan and/or additional deployments in a country, it may be helpful to launch funding for broadband including consideration of more local data centres, supported by reliable and the IoT to accelerate the growth of the IoT and to quality electricity, tailored tax incentives, and low or capitalize on these benefits . In the UK for instance, more flexible labour costs . Prime Minister David Cameron announced [in March 2014] an additional GBP 45 million funding for the 5 . Promote standards that facilitate interoperability development of IoT 1. Malaysia’s Ministry of Science, across the IoT ecosystem, foster investment, Technology and Innovation and its applied research competition and scale to enable cost-effective agency released a National Internet of Things (IoT) solutions . Strategic Roadmap in mid-2015 .2 Independently, 6 . Trust and confidence in the IoT are fundamental and some donor partners in the ICT4D community have must be designed into the IoT from the outset . Two also established “Principles for Digital Development” key components to ensure trust and confidence that aim to facilitate the adoption of the IoT in are privacy and security: a) Strategies to protect development .3 Governments and policy-makers privacy must take a range of risks into account from should work closely with industry to understand the a variety of different sources as well as adapt to local issues involved . This policy should also consider regulations; and b) Accelerate research into IoT- how new IoT systems can interface with pre-existing related security threats to minimize the downsides legacy infrastructure to protect and make full use of of the IoT across M2M and M2P communications . existing investments in infrastructure . The exponential growth of increased attack vectors 2 . Give consideration to sourcing any additional (in terms of type of data generated and the variety spectrum which may be needed for IoT . As of things and devices connected to the Internet) may discussed in Section 2, connectivity and therefore give attackers easy ways to access networked data . spectrum is a key part of supporting the expansion of This report has made the case for the use of IoT to improve people’s lives around the world . It has outlined 1 “‘Internet of things’ to get £45m funding boost”, 9 March 2014, available at: http://www .bbc com/news/business-26504696. a number of practical and technical considerations in 2 https://www telegeography. com/products/commsupdate/arti. - the deployment of IoT systems in the hope of improving cles/2015/07/14/malaysia-publishes-national-roadmap-for-iot/?utm_ outcomes in development projects and accelerating source=CommsUpdate&utm_campaign=9add5560b9-CommsUp- date+14+July+2015&utm_medium=email&utm_term=0_0688983330- the adoption of IoT in developing countries to improve 9add5560b9-11619241 people’s lives . 3 http://digitalprinciples org/.

47 Harnessing the Internet of Things for Global Development Annexes

Harnessing the Internet of Things for Global Development 48 Annex 1: Projects By Sector managers have access to daily summaries of the refrigerators’ temperatures, which are posted online to This Annex presents a glossary of additional information the Smart Connect website . about projects mentioned in this report . http://homes cs. washington. edu/~eorourke/papers/smart_. Health connect_nsdr .pdf

Nexleaf Analytics — Cold Chain Monitoring STAMP2

Nexleaf Analytics is a non-profit technology company The United States Agency for International that builds wirelessly connected devices and sensor Development (USAID) has supported and employed IoT technologies for critical public health and environmental solutions via connected wearable technologies, most interventions . Nexleaf’s Cold Trace remote temperature recently through its backing of the Sensor Technology monitor uses mobile phones to collect and wirelessly and Analytics to Monitor, Predict, and Protect Ebola transmit temperature data from refrigerated units that Patients program (STAMP2 for short), created by the store vaccines . Sensors wirelessly upload temperature Scripps Translational Science Institute (STSI) . This data and the system generates SMS and email alerts intervention has been tested on Ebola patients in the when vaccine refridgeration equipment registers United States and is being scaled up to meet the dangerous temperatures . By providing real-time needs of other government agencies for the Ebola information on equipment performance, vaccine supply treatment strategy in Liberia . STAMP2 collects patient chains become more secure . Fridge performance data, including ECG, heart rate, oxygen saturation, information by model and power availability data by body temperature, respiratory rate and position . location can improve forecasting and capacity planning This data is then sent to a centralized platform so it in the future . can be monitored and analyzed over a long period http://nexleaf org/technology/cold-chain-monitor. of time to alert physicians to abnormal changes in a patient’s behavior or health . The STAMP2 sensor will Smart Connect – Cold Chain Monitoring be deployed in Ebola-stricken areas using a connected Smart Connect is a “communication appliance” health patch or “Smart Band-Aid” . The fully equipped developed by PATH and Inveneo that uses SMS to sensor-enabled band-aid is estimated to cost improve the reliability and performance of one of approximately US$100 with a maximum battery life of the most important systems in all of global health: 10 days – making it ideal for use in field-based Ebola the medical “cold chain” . With over 200,000 vaccine treatment centers . Deployment of the STAMP2 sensor refrigerators in use in the developing world alone, most would improve the broader Ebola response initiative by of them in harsh and remote environments, keeping reducing emergency response times in critical areas the cold chain up and running is a major challenge that and enabling emergency responders to detect Ebola Smart Connect seeks to overcome . The device was patients earlier and to monitor them more efficiently . developed to confront communication barriers and http://mobihealthnews com/40564/scripps-wins-usaid-grant-. address the cold chain challenge by bringing a “digital to-monitor-ebola-patients-with-medical-wearables/ dial tone” to remote health posts in the developing world . Representing a significant departure from the traditional method of tracking and recording refrigerator temperatures, Smart Connect now records the temperature and sends out alert messages whenever there is a problem . Messages are then sent to a website and automatically relayed to service technicians who can respond immediately . Additionally,

49 Harnessing the Internet of Things for Global Development Water & Sanitation higher than the data captured by the motion sensors, suggesting over-reporting in the verbal and self- The Hidrosonico Project reported surveys .

Water flow sensors are being used to aid in the http://www .mercycorps org/sites/default/files/Instrumented%20. collection of hydrological data in developing countries Monitoring%20Indonesia .pdf where local data on river flow and levels may not be regularly collected . These sensors can also help SmartPump (Mobile Enabled Transmitter) - Maintenance with early warning for flooding . The Hidrosónico is a Service Provider Model water stream gauge that uses sonar range sensors The Smith School of Enterprise and the Environment to measure the distance to water surface level . at Oxford University, in conjunction with the UK The module sends the readings on regular basis to Department of International Development, conducted recipients (via SMS or email) or to a cloud application . a study in the Kyuso District of Kitui County, Kenya, The unit is also equipped with a rain gauge to monitor from August 2012 until December 2013, in which precipitation, and is currently deployed in Honduras . water pumps were equipped with GSM transmitters http://dai com/our-work/solutions/dai-maker-lab;. that sent data via SMS over existing mobile phone https://github com/DAI-Maker-Lab/hidrosonico. networks . Following a series of proof-of-concept tests in Lusaka (July 2011), transmitters were installed RW Siaga Plus+ Program in 21 handpumps, that offered: (1) measurement of handpump usage and associated volumetric water use; Between September 2009 and September 2011, (2) remote surveillance of maintenance service delivery Mercy Corps in Indonesia conducted and evaluated and downtime; and (3) objective data to improve various approaches to their water and sanitation infrastructure planning and investment, and promote programme RW Siaga Plus+, in the Indonesian sector accountability . Data transmitted from pumps are kelurahan (sub-district) of Margahayu, Bekasi, as well captured in a relational database and presented using a as in sixteen poor urban neighborhoods in three other bespoke graphic user interface . The Ministry of Water kelurahan located throughout West Jakarta . A major and Irrigation, the Water Services Regulatory Board goal of this programme was the creation of healthy and the District Water Office were consulted for their physical environments in urban poor settlements inputs . The results of the study are compelling: 98% of through increased access to clean water supplies and handpumps now work in Kyuso (compared to a rough improved sanitation . Using two different evaluation estimate of 70% in Africa), five times more revenue is methods, the organization and its partners measured being collected, handpump downtime reduced 10-fold the programme’s effectiveness on achieving behavioral from 27 days to less than 3 days, and now 80% of change and overall programme targets, enabled by users are willing to prepay after the trial (compared to data collected from water flow sensors deployed in the less than 20% before the trial) . field . In one programme delivering water, hygiene and http://www .smithschool ox. .ac .uk/library/reports/SSEE_Rights to sanitation interventions, flow sensors, in combination Results_FINAL_March2014 pdf. with motion detectors, amassed data that was used to measure the impact of behavior change training, with SWEETSense - CellPump the intent of increasing hygienic actions (washing after latrine use) . The study found community participants SweetSense Inc . provides low-cost remote monitoring were washing hands after latrine usage; however, the solutions for water, energy, and environmental projects survey responses (those who indicated they wash to reduce client operating costs, improve technical their hands after using the latrine) were significantly performance, and increase overall effectiveness . The

Harnessing the Internet of Things for Global Development 50 social enterprise corporation, in collaboration with Agriculture & Livelihoods Portland State University’s SweetLab, has developed CellPump, a smart sensor that can be installed within Kilimo Salama - Connected Weather Station the pump head of Afridev and India Mark 2 water Beginning in 2009 with a pilot project offering index pumps . Water flow data are transmitted by sensors insurance to 200 farmers in Kenya, the Syngenta over GSM cellular networks (98% of the country Foundation’s Kilimo Salama (“Safe Farming”) weather has cell coverage) and monitored by Living Water index insurance programme has helped over 51,000 International via an online dashboard . Currently over farmers in Kenya and 14,000 farmers in Rwanda 200 sensors have been provided for Living Water to date . The programme’s solar-powered weather International, deployed in various remote villages in stations collect weather data every 15 minutes, which Rwanda . The project is supported in part by the GSM are then aggregated and compared to historical Association and the UK Department for International weather data at the end of each growing season . Development, and has been implemented in Any payout owed is calculated and sent to farmers partnership with MTN, the Rwanda Ministry of Natural via mobile phone . Syngenta, on behalf of Kilimo Resources, and other local government authorities . Salama, partnered with Safaricom in 2010, providing http://www .sweetsensors com/applications/cellpump/. a less expensive and dense communications network for product sales and customer communication . WellDone - MoMo Using Safaricom’s M-PESA mobile banking system, WellDone builds technological tools that empower the programme keeps index insurance premiums resource-constrained communities with the data affordable for smallholder farmers, who receive their they need to invest in and maintain lasting critical index insurance policy numbers and premium receipts infrastructure . The organization works at the local via SMS . Payouts are sent electronically through level to ensure that these communities become M-PESA as well, which has reduced the associated independent of development assistance, and focuses risks of adverse weather and provided farmers with on the distribution of technology that facilitates more a financial safety net . Subsequently, the regions that reliable and robust infrastructure development . Notably, have participated in the Kilimo Salama programme MoMo (mobile monitor) is a mobile device produced have enjoyed increased agricultural investment and by WellDone that collects data to track the progress improved wellbeing, while partnering microfinance of infrastructure projects, report on their progress, institutions have seen their loan portfolios increase in and notify appropriate personnel when maintenance value, sometimes by as much as double within the first is required . MoMo has been used in handpumps, six months . measuring their functionality, their frequency of use, http://www .ifc org/wps/wcm/connect/. and their hourly water flow . The data generated by e0ed35804c33fc309479def12db12449/KS+story . remote MoMos are then aggregated using WellDone’s pdf?MOD=AJPERES Smarter Villages software, which alerts local repair teams by SMS whenever a well in their area breaks Nano Ganesh – Micro Irrigation Solution down . The software also displays trends in water use, In India, the Nano Ganesh solution is a low-cost which can be used at the local, and potentially national, product set to provide small-shareholder farmers with level to better allocate resources and plan projects a tool that can remotely control their micro irrigation according to levels of need . pumps . In India, about 25 million water pumps are in https://welldone org. use for farm irrigation . Many of these pumps have to be manually operated, based on rainwater conditions, electricity availability and crop needs . For the average small-scale farmer, the variability of these factors on a

51 Harnessing the Internet of Things for Global Development day-to-day basis adds extra burdens in terms of time, partnership introduced low-cost personal location labor and fuel costs . In many cases, farmers need to beacons (PLBs) as part of a community-based system travel long distances and/or through difficult conditions for identifying illegal, unreported and unregulated to access their pumps from their households . The fishing . Now, when local fishermen identify illegal Nano Ganesh unit works by attaching to the irrigation fishing boats in the Timor-Leste waters, the location pump, and serving as an actuator, which can turn on beacon is activated and data are sent to an Internet- and off via basic commands from the farmer’s simple enabled mapping platform, which authorities utilize feature phone (2G) . The farmer is also able to check to track and apprehend illegal fishing boats or initiate the availability of electricity at the pump, as well as the rescue services . availability of water near the pump (with an additional http://www .fao org/3/a-i4622e. .pdf water sensor) . As of August 2014, twenty thousand farmers in India are benefiting from Nano Ganesh . Resiliency, Climate Change, & Pollution Mitigation http://www .fao org/3/a-i4622e. .pdf Fresh Air in Benin

Sensors for Tea Plantations Urban air pollution is a major problem in many The quality of tea, and its resulting sales price, are developing country cities . The World Health determined in part by water composition of its soil Organization (WHO) attributes one in eight deaths and surrounding environment . As a major export of Sri worldwide to polluted air, as dirty air causes lung Lanka and Rwanda, it is essential that moisture levels damage, heart disease, strokes and cancer . The WHO be monitored and recorded; however, traditional manual also estimates that indoor air pollution in homes in methods have proven inefficient and time-consuming . Africa contributed to nearly 600,000 deaths in 2012 . Plantations in both countries utilize WSNs to monitor To measure the extent of the problem, air quality moisture, pH levels, carbon, nitrogen, potassium, calcium sensors are being deployed in a range of cities to track and magnesium levels of the soil in which their tea levels and changes in pollutants . One such project, grows . Sensors and connectivity modules are powered Fresh Air in Benin, is focused on developing a network by solar panels, and the data that they capture are of air quality sensors to capture and send data every transmitted wirelessly . At a relatively low cost, sensors 20 minutes via GSM connectivity . deployed in the field provide actionable feedback every https://docs .google com/folderview?id=0B_. fifteen seconds, saving time, money, and crops . lA8alw74UiVGJuU0hGV01scVE&usp=docslist_ http://wireless .ictp .it/rwanda_2015/ api&tid=0B81LFtxt5uZKeEphNG9PRTFPR2c http://www .ictp .it/about-ictp/media-centre/news/2015/6/ Smart Fire Sensors for Slums & Informal Settlements teatime-with-iot .aspx A regularly occurring tragedy in high-density urban slums, fires are particularly deadly given the rapid http://wireless .ictp .it/school_2015/presentations/CaseStudies/ and haphazard nature of these neighborhoods’ IoTforTeaPlantation-SriLanka-Minuri .pdf development, including narrow doorways that are Timor-Leste’s Community-Based IUU Reporting System often blocked, high population density, and the lack of adequate first response . The Red Cross has proposed In Timor-Leste, the National Directorate of Fisheries that local entrepreneurs help solve the problem by and Aquaculture (NDFA), working in partnership installing connected fire alarm systems across slums with the FAO-Spain Regional Fisheries Livelihoods and other informal settlements to quickly notify Programme for South and Southeast Asia (RFLP), residents when a fire first starts . The organization introduced a plan using radio location beacons to has also explored the option of developing low-cost, protect local waters from illegal fishing and to provide solar-powered sensors networked together to quickly emergency response to fishermen in distress . The detect when a fire has broken out and notify the

Harnessing the Internet of Things for Global Development 52 dwelling’s resident(s), as well as surrounding neighbors . The technology includes an in-vehicle sensor monitor Upon detection of smoke, the network sounds an that reports on vehicle velocity, acceleration, heavy- alarm and communicates the threat to other homes braking, and transmits data via GSM . A web application nearby through SMS or other modalities . Additionally, collects and analyzes each device’s data streams, the networked sensors are GPS-enabled and have the launching notifications to stakeholders (driver, owner, ability to identify a fire’s point of origin and transmit fleet manager, municipality) . An Android application, that information to authorities so that they may more coupled with SMS alerts, gives owners access to accurately target fire extinguishment efforts . Currently, detailed real-time information about each vehicle the intervention is being tested in Nairobi and Cape in their fleet . Through the application, owners (and Town, with participation by two thousand households . other stakeholders) can drill down into information on https://drive .google com/file/. locations, routes, and recent productivity and safety d/0B1vf6TLGIC0yZk9UU2t2UGFmNlE/view?usp=sharing events and take action against unsafe drivers . https://www echomobile. org. South Indian Ocean Tsunami Warning System http://pedl cepr. org/sites/default/files/Research%20Note_. Following the 2004 Indian Ocean tsunami that Mitigating%20Market%20Frictions%20by%20Monitoring%20 devastated coastal areas in India, Sri Lanka, Thailand SME%20Employees_3022 .pdf and Indonesia, the international community united to establish the Indian Ocean Tsunami Warning System Natural Resource Management whereby kinetic sensors (measuring waves and water Acoustic Sensors for Seabird Monitoring & Identification flow) placed on the ocean floor communicate data on potential tsunamis to disk buoys floating on the ocean Funded with a grant from the National Science surface via acoustic telemetry . Once collected, buoys Foundation, Nexleaf Analytics developed and tested then upload the information to government authorities a low-cost tool to remotely monitor seabirds; test via satellite connectivity . deployments focused on colonies located in the Pacific . From December 2011 through May 2012, http://www .kophuket com/phuket/homemenu/tsunami. html. the organization worked with the U S. . Fish & Wildlife http://iotic .ioc-unesco org/indian-ocean-tsunami-warning-. Service and the Coastal Conservation & Action Lab system/tsunami-early-warning-centres/56/national-tsunami- at the University of California at Santa Cruz to deploy warning-centres four sensors on Tern Island, located in the French Frigate Shoals in the North-West Hawaiian islands . The Echo Mobile - Fleet Management for Public Safety deployment consisted of four wildlife acoustic monitors In Kenya, a large share of road traffic accidents occur and two WiFi radio repeaters powered by 20W solar in the semi-formal public transport sector using panels, and one gateway computer . An estimated minibuses (known as matatus) . With a third of the 400 MB of data were generated daily . The project population relying on informal transit, unsafe driving demonstrated the performance capabilities of satellite- practices are a substantial cause of preventable death . based WSNs for long-term monitoring of seabird The challenge is that mini-bus owners lease buses colonies, a method which may prove useful for other daily to their drivers but cannot observe the drivers’ projects dedicated to conservation efforts or natural behavior . As most drivers are paid for each route they resource management . complete, they may adopt reckless habits: speeding, http://nexleaf org/project/tern-island. weaving between lanes, and driving on sidewalks . Misaligned incentives lead to an unsafe transport system particularly for passengers . Echo Mobile has developed a sensor-based system for vehicle owners to track where and how their vehicles are being driven .

53 Harnessing the Internet of Things for Global Development Anti-Poaching Interventions photos from remote camera traps . Wadi Drones further eliminate the need to employ a costly helicopter to reach Illegal poaching of large wildlife is a major concern camera traps during the summer months, when high across Sub-Saharan Africa . In 2014 alone, 40,000 temperatures make hiking conditions dangerous . elephants were illegally killed, primarily for their tusks, which are highly sought after in the ivory trade . From http://wadi .io 26 million elephants in 1800 to fewer than one million Rainforest Connection today, the African elephant is at risk of extinction by illegal poaching, but it is not alone . Demand for Rainforest Connection transforms recycled smartphones the horns of the black rhinoceros has lead to 96% into autonomous, solar-powered listening devices decline in the species’ population from 1970 to that can pinpoint signs of destructive activity at great 1992, with fewer than 20,000 animals remaining . distance . It is a scalable, real-time logging detection To combat these and other poaching operations, system, pinpointing deforestation activity as it occurs . a wide number of projects have commissioned With an initial field project only located in Sumatra, emerging IoT technologies, such as connected drones Rainforest Connection (RFCx) is now expanding to three for surveillance, as well as long-range wide area more endangered areas in the rainforests of Indonesia technologies, to track wildlife and to monitor activity at the Amazon, and Africa . By partnering with local the boundaries of game parks . communities, indigenous groups, and organizations that http://airshepherd org/. are committed to responding to real-time interventions, the organization hopes to demonstrate that their system http://gblogs cisco. com/uki/week-4-how-can-technology-. of next-generation devices can operate on a global help-the-anti-poaching-activities/ scale, in any forest, anywhere . https://rfcx org. Wadi Drone

The Wadi Drone is a fixed wing airplane with a 2 .5​ metre wingspan carrying a small communications payload that retrieves information from ground-based scientific measurement devices . Comprising four NYU Abu Dhabi students, the Wadi Drone development team collaborated with the Emirates Wildlife Society (WWF) and the country’s first national park, Wadi Wurayah National Park located in Fujairah, to craft and deploy the drone, which was a 2015 finalist in the UAE Drones for Good Award competition . Within the parameters of the national park, the drone flies over mountains and through valleys to wirelessly download photographs taken by ground-based camera traps that automatically capture images of wildlife as they pass in front of the camera’s motion sensor . Collecting data via permanent communications infrastructure could interfere with the natural environment or endanger workers, and the Wadi Drone project assists the conservation efforts of the Emirates Wildlife Society by both increasing the rate at which photographic data of wildlife can be analyzed by experts, and by reducing the human risks associated with the current method of hiking to retrieve

Harnessing the Internet of Things for Global Development 54 Energy an enumerator) to determine whether the latter method was accurately reflecting usage captured by sensors . M-KOPA Solar In this case, the study found that survey participants M-Kopa, a pay-as-you-go Energy Service Company were over-reporting daily cooking hours by 1 .2 hours (ESCO) for off-grid customers in Kenya, leverages on average, and daily cooking events by 1 3. events . machine-to-machine (M2M) technology to fulfill The results suggest that survey-based methods for its mission of providing high-quality energy at an monitoring and evaluation may be misleading in their affordable rate to everyone . Their solar home system presentation of actual impact and usage . According from D .Light Design is comprised of photovoltaic cells, to the results of this study, data collected by sensors a battery system and a communications model, which therefore represents the more reliable method . individuals purchase altogether at a discount with http://www .state .gov/r/pa/prs/ps/2015/09/247240 .htm capital costs amortized over a designated purchase Wilson, Daniel L ,. et . al . “Comparing Cookstove Usage period . After the system is installed, customers are able Measured with Sensors Versus Cell Phone-Based to utilize the electricity generated by the solar cells to Surveys in Darfu, Sudan .” Technologies for Development: power home appliances . Following an initial deposit, What is Essential . they must make regular payments (usually via mobile money systems, e .g . M-Pesa or Airtel Money) to continue using the device . M-Kopa remotely monitors the amount of electricity captured, the allocation that is stored, and, through the device’s connectivity module, if the device is working appropriately . http://www .m-kopa com.

Monitoring & Evaluation of Cookstoves

Across the developing world, wood and charcoal- burning cookstoves are used extensively to prepare meals and to provide a source of heat inside homes . The resulting indoor air pollution contributes to approximately 4 million deaths a year, out of the 3 billion people worldwide who utilize biomass to prepare their meals . In response, a number of initiatives are in place to help lower-income households transition to less injurious methods of meal preparation and heating by way of improved cookstove projects . The U .S . Government has supported a five-year initiative of the “Global Alliance for Clean Cookstoves” which aims to achieve a goal of enabling 100 million homes to adopt clean and efficient cooking solutions by 2020 . IoT sensors play a major role in this initiative by helping to measure in real time the black carbon emitted by cookstoves, and also in the monitoring and evaluation of projects that disseminate upgraded cookstoves . In one particular project in the Sudan, the use of improved cookstoves via sensor-enabled recording instruments was compared to traditional survey data (captured by

55 Harnessing the Internet of Things for Global Development Annex 2: Different Characteristics of Wireless IoT Connectivity Technologies licensed licensed licensed licensed unlicensed unlicensed unlicensed unlicensed unlicensed unlicensed unlicensed unlicensed unlicensed unlicensed unlicensed Spectrum License

-

Europe; Europe; 2 . 4 GHz 2 . 4 GHz 2 . 4 GHz licensed licensed 433 Mhz Various - Various Various - Various 13 . 56 MHz Frequency 2 . 4GHz/5GHz less-W: TVWS less-W: 900 MHz in US) 2 . 4GHz/ 900Mhz 2 . 5 GHz; 3 GHz Global GSM bands USA (902-928 MHz) USA (868 MHz in Europe; (868 MHz in Europe; Spectrum/ Operating (915 MHz, 868 MHz) spectrum but WiMAX Sub-1 GHz ISM bands – Korea (917-923 . 5 MHz);Korea ISM bands (868 MHz in Europe (863-868 . 6 MHz);Europe 900 MHz in US); Weight Weightless-N: ISM bands Weightless-N: 433 MHz, ISM bands (868 120-150 kHz; 12 . 56 MHz, No uniform global licensed forum published 3 licensed spectrum profiles: 2 . 3 GHz; spectrum profiles: MHz, 900 MHz), 2 . 5-5 8 GHz Japan (950 . 8 MHz – 957 6z MHz); <$1 $10+ Active: Active: $1-$15 <$1-$5 $5-$25 Passive: Passive: $1- $15 $1 - $15 $35-$50 $1 - $15 $1 - $15 $1 - $15 $1 - $15 $1 - $15 $1 - $15 $80-$120 Module Cost

n/a n/a 4-8 Days (idle) years years Up to Up to (idle), (com) (com) Hours (com) Hours (Battery) 10 years 10 years 50 hours up to two two up to 4-8 hours 2-4 hours 120 hours 2-3 Hours hours(com) Active Tags: Tags: Active 10-20 years 10-20 years Up to 4 years Up to Passive Tags: Tags: Passive 12 days (idle) 12 days 36 days (idle) 36 days 20 days (idle) 20 days communicating Operating Life 1 Mbps 24 Mbps 200Kbps 200Kbps 250 Kbps 200 Kbps 200 Kbps 424 Kbps 100 Kbps Throughput 34 Mbps – 1 Gbps 1Gbps (802 . 11ac) 3 Mbps – 100 11 Mbps (802 . 11b); 384 Kbps – 10 Mbps 9 . 6 Kbps – 384 100 kbps (802 . 11ah) Max Bandwidth/ Data 250 Mbps (802 . 11n); 54 Mbps (802 . 11a/g); (30 50m 30m 2 km 10cm 300m 200m miles) 35 km 40 km 2-10 km 2-10 km Max Range up to 100km up to up to 100km up to up to 1000m up to Active: 100m Active: Passive: 10m Passive: 10-100 meters NFC RFID LoRa Wi-Fi ANT+ WiMax Dash 7 ZigBee 802 . 15 4g Weightless Cellular 4G/ LTE Bluetooth 4 . 0 LE Bluetooth Wi-Fi (802 . 11ah) Technology Name Technology 3G (UTMS, HSPA) 3G (UTMS, 2G (GSM, GPRS, EDGE) 2G (GSM, GPRS, Type WLAN WPAN WWAN Wireless Wireless Technology Technology

Note: Non-exhaustive . Source: Cisco Systems .

Harnessing the Internet of Things for Global Development 56 Annex 3: Sample Sensor Prices (Retail)

Price Sensor Name (rounded up) Integrated WiFi High Temperature Sensor $204 Integrated WiFi Humidity Sensor $180 Dual-Range Force Sensor $110 Ultrasonic Range Finder; CO2 Sensor $100 Gas Pressure Sensor $83 Vernier Motion Detector $75 Oxygen Sensor $60 RFID Starter Kit; Voice Recognition Shield; Color Detector Sensor; JPEG Color Camera $50 TTL Interface; Soil Temperature and Moisture Sensor Wind Speed Sensor $45 Humidity & Temperature Sensor $42 Multichannel Gas Sensor; Liquid Level Sensor; Soil Moisture and Temperature; Liquid Level Sensor; $40 GPS Breakout Sensor; Wearable GPS Module G5 Water Flow Sensor; RFID Reader; RFID Sensor Module; Ultrasonic Range Finder Lite; $30 Temperature and Humidity Sensor Board; Distance and Gesture Sensor; Liquid Flow Meter Nominal Thread; Ultrasonic Rangefinder; $25 Infrared Distance Sensor AttoPilot Voltage and Current; Pressure Sensor (up to 100 lbs); Pressure Sensor; High Accuracy $20 Barometer; Temperature and Humidity Sensor Humidity Sensor $17

G1 Water Flow Sensor; Camera Module; Dust Sensor; Weight Sensor 400kg $16 Ultrasonic Range Measurement Module; Barometric Sensor; Infrared Proximity Sensor; RGB and $15 Gesture Sensor; Temperature and Humidity Sensor; Triple-axis accelerometer; IR Distance Sensor Low Accuracy Barometer; Grove Gesture Sensor $14

Coulomb Counter; UV Sensor $13 AC Current Sensor high amperage $12 Sound Detector $11 Capacitive Touch Sensor; Load Sensor (up to 50kg); Air Quality Sensor; FM Receiver; Temperature and Humidity Probe; Barometric and Temperature Sensor; Altitude Sensor; Digital UV Index/IR/Visible $10 Light Sensor; Proximity Light Sensor; Liquid Flow Meter; PIR Motion Sensor Current Sensor; PIR Motion Sensor; Collision Sensor; Temperature and Humidity HP Sensor $9 RGP Light Sensor; Alcohol Sensor; RGB Color Sensor and IR Filter $8 Infrared shooting sensor; AC Current Sensor; HDR Digital Light Sensor; Microphone Amplifier $7 Luminosity Sensor; Loudness Sensor $6 Infrared reflective sensor; RFID Capsule; Ambient Light Sensor; Moisture Sensor; Sound Sensor $5 Vibration Sensor; Water Sensor $3

Source: Website review of sensors prices August 6 – 11, 2015 from Sparkfun, Seeed Technology Inc ., Adafruit, and Monnit .

57 Harnessing the Internet of Things for Global Development List of Acronyms and Abbreviations

API Application Programming Interface

FDA Federal Drugs Agency (of the United States)

GSMA Global System for Mobile Association

IoE Internet of Everything

IoT Internet of Things

ITU International Telecommunication Union

M2M Machine-to-Machine

M2P Machine-to-Person

MDGs Millennium Development Goals

NFC Near Field Communications

P2P Person-to-Person

PLB Personal Localization Beacon

SDGs Sustainable Development Goals

UN United Nations

US United States

US$ United States Dollar

WHO World Health Organization

WLAN Wireless Local Area Network

WSN Wireless Sensor Network

WWAN Wireless Wide Area Network

4G Fourth-generation mobile

5G Fifth-generation mobile

Harnessing the Internet of Things for Global Development 58 International Telecommunication Union Place des Nations CH-1211 Geneva 20 Switzerland

ISBN 978-92-61-16401-0

9 7 8 9 2 6 1 1 6 4 0 1 0

Geneva, 2016